Description:FORM 2

THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENT RULES, 2003

COMPLETE SPECIFICATION
(See Section 10 and Rule 13)

Title of invention:
METHODS AND SYSTEMS FOR GENERATING CAUSE-EFFECT SENTIMENT-ENRICHED KNOWLEDGE GRAPH FOR PORTFOLIO OPTIMIZATION

Applicant

Tata Consultancy Services Limited
A company Incorporated in India under the Companies Act, 1956
Having address:
Nirmal Building, 9th floor,
Nariman point, Mumbai 400021,
Maharashtra, India

Preamble to the description:
The following specification particularly describes the invention and the manner in which it is to be performed.
TECHNICAL FIELD
[001] The disclosure herein generally relates to the field of portfolio optimization, and more specifically to methods and systems for generating a cause-effect sentiment-enriched knowledge graph for portfolio optimization.

BACKGROUND
[002] Portfolio optimization is a prime business approach for any organization. The portfolio optimization helps to make right investment decisions across multiple financial options or assets. Further, the portfolio optimization helps the investor to receive a right guidance from various financial strategies, for a given level of risk and at the same time to yield maximum possible returns. Further, large number of the investors are their turning attention to prediction of stocks and trying to find different and more efficient ways of speculating about markets through the application of behavioral finance. Use of a knowledge graph is one of the technique which can be reasoned over rules to obtain the useful insights of the business and the market trends. Most of the conventional techniques in the art for the portfolio optimization use natural language processing techniques and text mining approaches for building the knowledge graphs. The development of the knowledge graph in natural language processing (NLP), researchers in the financial field began paying attention to the text mining in financial news.
[003] However conventional knowledge graphs, in particular the financial knowledge graphs are, incomplete, not accurate and not efficient to predict the market trends and stock exchanges, and so on, for various reasons. Some of the reasons include factors such as sentimental news, numerical relations present in the news updates, events such as war, occurrence of events, are technically challenging to include in the knowledge graph, due to the inter-relationship between them. Due to these technical challenges, the financial knowledge goes missing from the knowledge graph and results inaccurate and inefficient business solutions.

SUMMARY
[004] Embodiments of the present disclosure present technological improvements as solutions to one or more of the above-mentioned technical problems recognized by the inventors in conventional systems.
[005] In an aspect, a processor-implemented method for generating a cause-effect sentiment-enriched knowledge graph for portfolio optimization is provided. The method including the steps of: obtaining one or more macroeconomic variables, one or more impacts for each of the one or more macroeconomic variables, and a change in a stock; obtaining a definition associated with each of the one or more macroeconomic variables, using financial ontologies and a sentiment dictionary; obtaining a time-stamped financial trend information associated with each of the one or more macroeconomic variables, using one or more geography economy sources; determining one or more triplets from: (i) the one or more macroeconomic variables and (ii) the one or more impacts for each of the one or more macroeconomic variables, using a natural language processing (NLP) technique, wherein each triplet of the one or more triplets comprises at least one of either: (i) a macroeconomic variable of the one or more macroeconomic variables, or (ii) an impact of the one or more impacts; building a base knowledge graph comprising a plurality of nodes and one or more edges, based on the one or more triplets, using a graph database and the natural language processing (NLP) technique, wherein each node of the plurality of nodes is represented by either a macroeconomic variable or an impact present in each of the one or more triplets, and each edge of the one or more edges is represented by a relation between the two nodes of the plurality of nodes; determining, (i) one or more numerical relations associated with each triplet, (ii) one or more numerical sentiments associated with each triplet, (iii) one or more events and cooccurrence of the one or more events associated with each triplet, (iv) one or more company names and associated sustainability rankings associated with each triplet, and (v) cooccurrence of each microeconomic variable present in each triplet; updating, the base knowledge graph with (i) the one or more numerical relations associated with each triplet, (ii) the one or more numerical sentiments associated with each triplet, (iii) the one or more events and cooccurrence of the one or more events associated with each triplet, (iv) one or more company names and associated sustainability rankings associated with each triplet, (v) the cooccurrence of each microeconomic variable present in each triplet, and (vi) the change in the stock, to obtain the cause-effect sentiment-enriched knowledge graph; and performing a reasoning on the cause-effect sentiment-enriched knowledge graph, using a rule engine, to extract financial insights and patterns for the portfolio management optimization.
[006] In another aspect, a system for generating a cause-effect sentiment-enriched knowledge graph for portfolio optimization is provided. The system includes: a memory storing instructions; one or more Input/Output (I/O) interfaces; and one or more hardware processors coupled to the memory via the one or more I/O interfaces, wherein the one or more hardware processors are configured by the instructions to: obtain one or more macroeconomic variables, one or more impacts for each of the one or more macroeconomic variables, and a change in a stock; obtain a definition associated with each of the one or more macroeconomic variables, using financial ontologies and a sentiment dictionary; obtain a time-stamped financial trend information associated with each of the one or more macroeconomic variables, using one or more geography economy sources; determine one or more triplets from: (i) the one or more macroeconomic variables and (ii) the one or more impacts for each of the one or more macroeconomic variables, using a natural language processing (NLP) technique, wherein each triplet of the one or more triplets comprises at least one of either: (i) a macroeconomic variable of the one or more macroeconomic variables, or (ii) an impact of the one or more impacts; build a base knowledge graph comprising a plurality of nodes and one or more edges, based on the one or more triplets, using a graph database and the natural language processing (NLP) technique, wherein each node of the plurality of nodes is represented by either a macroeconomic variable or an impact present in each of the one or more triplets, and each edge of the one or more edges is represented by a relation between the two nodes of the plurality of nodes; determine (i) one or more numerical relations associated with each triplet, (ii) one or more numerical sentiments associated with each triplet, (iii) one or more events and cooccurrence of the one or more events associated with each triplet, (iv) one or more company names and associated sustainability rankings associated with each triplet, and (v) cooccurrence of each microeconomic variable present in each triplet; update the base knowledge graph with (i) the one or more numerical relations associated with each triplet, (ii) the one or more numerical sentiments associated with each triplet, (iii) the one or more events and cooccurrence of the one or more events associated with each triplet, (iv) one or more company names and associated sustainability rankings associated with each triplet, (v) the cooccurrence of each microeconomic variable present in each triplet, and (vi) the change in the stock, to obtain the cause-effect sentiment-enriched knowledge graph; and perform a reasoning on the cause-effect sentiment-enriched knowledge graph, using a rule engine, to extract financial insights and patterns for the portfolio management optimization.
[007] In yet another aspect, there is provided a computer program product comprising a non-transitory computer readable medium having a computer readable program embodied therein, wherein the computer readable program, when executed on a computing device, causes the computing device to: obtain one or more macroeconomic variables, one or more impacts for each of the one or more macroeconomic variables, and a change in a stock; obtain a definition associated with each of the one or more macroeconomic variables, using financial ontologies and a sentiment dictionary; obtain a time-stamped financial trend information associated with each of the one or more macroeconomic variables, using one or more geography economy sources; determine one or more triplets from: (i) the one or more macroeconomic variables and (ii) the one or more impacts for each of the one or more macroeconomic variables, using a natural language processing (NLP) technique, wherein each triplet of the one or more triplets comprises at least one of either: (i) a macroeconomic variable of the one or more macroeconomic variables, or (ii) an impact of the one or more impacts; build a base knowledge graph comprising a plurality of nodes and one or more edges, based on the one or more triplets, using a graph database and the natural language processing (NLP) technique, wherein each node of the plurality of nodes is represented by either a macroeconomic variable or an impact present in each of the one or more triplets, and each edge of the one or more edges is represented by a relation between the two nodes of the plurality of nodes; determine (i) one or more numerical relations associated with each triplet, (ii) one or more numerical sentiments associated with each triplet, (iii) one or more events and cooccurrence of the one or more events associated with each triplet, (iv) one or more company names and associated sustainability rankings associated with each triplet, and (v) cooccurrence of each microeconomic variable present in each triplet; update the base knowledge graph with (i) the one or more numerical relations associated with each triplet, (ii) the one or more numerical sentiments associated with each triplet, (iii) the one or more events and cooccurrence of the one or more events associated with each triplet, (iv) one or more company names and associated sustainability rankings associated with each triplet, (v) the cooccurrence of each microeconomic variable present in each triplet, and (vi) the change in the stock, to obtain the cause-effect sentiment-enriched knowledge graph; and perform a reasoning on the cause-effect sentiment-enriched knowledge graph, using a rule engine, to extract financial insights and patterns for the portfolio management optimization.
[008] In an embodiment, the one or more macroeconomic variables and the one or more impacts for each of the one or more macroeconomic variables, from one or more news articles, using the natural language processing (NLP) technique, wherein each of the one or more news articles comprises one or more financial statements; and the change in the stock, from stock return sources, using the natural language processing (NLP) technique.
[009] In an embodiment, the one or more numerical relations associated with each triplet, are determined by: extracting one or more financial entities present in the triplet, wherein the one or more financial entities are either the macroeconomic variable or the impact for the macroeconomic variable; annotating one or more connecting verbs associated with the one or more financial entities, using the financial ontologies; determining one or more sentiments for each financial entity of the one or more financial entities, using the NLP technique and the sentiment dictionary; refining a pre-trained word vector model, with the extracted one or more financial entities, the annotations of the connected verbs associated with the extracted one or more financial entities and the determined one or more sentiments for each financial entity of the one or more financial entities, to obtain the word vector model; and passing the extracted one or more financial entities, the annotations of the connected verbs associated with the extracted one or more financial entities and the determined one or more sentiments for each financial entity of the one or more financial entities, to the word vector model, to obtain the one or more numerical relations associated with each triplet.
[010] In an embodiment, the one or more numerical sentiments associated with each triplet, are determined using a trained neural network model, wherein the trained neural network model is obtained by: extracting one or more financial entities present in the triplet, wherein the one or more financial entities are either the macroeconomic variable or the impact for the macroeconomic variable; obtaining one or more embeddings from the one or more financial entities present in the triplet, using a word vector model; determining one or more sentiments for each financial entity of the one or more financial entities, using the NLP technique and the sentiment dictionary; obtaining the definition associated with each of the one or more financial entities present in the triplet, using the financial ontologies; assigning a label for each triplet, based on the one or more financial entities present in the triplet; assigning the determined one or more sentiments for each triplet, based on the label assigned for each triplet; and training a neural network model with (i) the one or more embeddings from the one or more financial entities present in the triplet, (ii) the one or more sentiments associated with the triplet, (iii) the definition associated with the one or more financial entities present in the triplet, (iv) the assigned label for each triplet, and (v) the assigned one or more sentiments for each triplet, to obtain the trained neural network model.
[011] In an embodiment, the one or more events and the cooccurrence of the one or more events associated with each triplet, are determined from (i) the one or more macroeconomic variables, and (ii) the one or more impacts for each of the one or more macroeconomic variables, using the NLP technique.
[012] In an embodiment, the one or more company names and the associated sustainability rankings associated with each triplet, are determined from annual reports, using the NLP technique.
[013] In an embodiment, the cooccurrence of each microeconomic variable present in each triplet, are determined from the cooccurrence of the one or more events associated with the one or more macroeconomic variables, using the NLP technique.
[014] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention, as claimed.

BRIEF DESCRIPTION OF THE DRAWINGS
[015] The accompanying drawings, which are incorporated in and constitute a part of this disclosure, illustrate exemplary embodiments and, together with the description, serve to explain the disclosed principles:
[016] FIG. 1 is an exemplary block diagram of a system for generating a cause-effect sentiment-enriched knowledge graph for portfolio optimization, in accordance with some embodiments of the present disclosure.
[017] FIGS. 2A and 2B illustrates exemplary flow diagrams of a processor-implemented method for generating the cause-effect sentiment-enriched knowledge graph for portfolio optimization, in accordance with some embodiments of the present disclosure.
[018] FIG. 3 shows an exemplary based knowledge graph build from the one or more triplets, in accordance with some embodiments of the present disclosure.
[019] FIG. 4 shows an exemplary cause-effect sentiment-enriched knowledge graph for portfolio optimization, in accordance with some embodiments of the present disclosure.

DETAILED DESCRIPTION OF EMBODIMENTS
[020] Exemplary embodiments are described with reference to the accompanying drawings. In the figures, the left-most digit(s) of a reference number identifies the figure in which the reference number first appears. Wherever convenient, the same reference numbers are used throughout the drawings to refer to the same or like parts. While examples and features of disclosed principles are described herein, modifications, adaptations, and other implementations are possible without departing from the scope of the disclosed embodiments.
[021] The present disclosure solves the technical problems in the art for by generating a cause-effect sentiment-enriched knowledge graph for portfolio optimization. The present disclosure combines the domain knowledge such as financial ontologies, microeconomic factors, and a business knowledge in picture considering the events happening around the words, sentimental relations, numerical relations associated with the sentimental factors and so on, along with the news for generating the cause-effect sentiment-enriched knowledge graph for portfolio optimization. The obtained cause-effect sentiment-enriched knowledge graph is complete and efficient for analyzing and predicting the market trends with their macroeconomic factors and their movement. The best optimized results are achieved from the obtained cause-effect sentiment-enriched knowledge graph, as the news information is also included to understand the market with the objective of effective portfolio optimization Further, the cause-effect sentiment-enriched knowledge graph can also be used to control the financial risks in the business activities.
[022] Referring now to the drawings, and more particularly to FIG. 1 through FIG. 4, where similar reference characters denote corresponding features consistently throughout the figures, there are shown preferred embodiments and these embodiments are described in the context of the following exemplary systems and/or methods.
[023] FIG. 1 is an exemplary block diagram of a system 100 for generating a cause-effect sentiment-enriched knowledge graph for portfolio optimization, in accordance with some embodiments of the present disclosure. In an embodiment, the system 100 includes or is otherwise in communication with one or more hardware processors 104, communication interface device(s) or input/output (I/O) interface(s) 106, and one or more data storage devices or memory 102 operatively coupled to the one or more hardware processors 104. The one or more hardware processors 104, the memory 102, and the I/O interface(s) 106 may be coupled to a system bus 108 or a similar mechanism.
[024] The I/O interface(s) 106 may include a variety of software and hardware interfaces, for example, a web interface, a graphical user interface, and the like. The I/O interface(s) 106 may include a variety of software and hardware interfaces, for example, interfaces for peripheral device(s), such as a keyboard, a mouse, an external memory, a plurality of sensor devices, a printer and the like. Further, the I/O interface(s) 106 may enable the system 100 to communicate with other devices, such as web servers and external databases.
[025] The I/O interface(s) 106 can facilitate multiple communications within a wide variety of networks and protocol types, including wired networks, for example, local area network (LAN), cable, etc., and wireless networks, such as Wireless LAN (WLAN), cellular, or satellite. For the purpose, the I/O interface(s) 106 may include one or more ports for connecting a number of computing systems with one another or to another server computer. Further, the I/O interface(s) 106 may include one or more ports for connecting a number of devices to one another or to another server.
[026] The one or more hardware processors 104 may be implemented as one or more microprocessors, microcomputers, microcontrollers, digital signal processors, central processing units, state machines, logic circuitries, and/or any devices that manipulate signals based on operational instructions. Among other capabilities, the one or more hardware processors 104 are configured to fetch and execute computer-readable instructions stored in the memory 102. In the context of the present disclosure, the expressions ‘processors’ and ‘hardware processors’ may be used interchangeably. In an embodiment, the system 100 can be implemented in a variety of computing systems, such as laptop computers, portable computers, notebooks, hand-held devices, workstations, mainframe computers, servers, a network cloud and the like.
[027] The memory 102 may include any computer-readable medium known in the art including, for example, volatile memory, such as static random access memory (SRAM) and dynamic random access memory (DRAM), and/or non-volatile memory, such as read only memory (ROM), erasable programmable ROM, flash memories, hard disks, optical disks, and magnetic tapes. In an embodiment, the memory 102 includes a plurality of modules 102a and a repository 102b for storing data processed, received, and generated by one or more of the plurality of modules 102a. The plurality of modules 102a may include routines, programs, objects, components, data structures, and so on, which perform particular tasks or implement particular abstract data types.
[028] The plurality of modules 102a may include programs or computer-readable instructions or coded instructions that supplement applications or functions performed by the system 100. The plurality of modules 102a may also be used as, signal processor(s), state machine(s), logic circuitries, and/or any other device or component that manipulates signals based on operational instructions. Further, the plurality of modules 102a can be used by hardware, by computer-readable instructions executed by the one or more hardware processors 104, or by a combination thereof. In an embodiment, the plurality of modules 102a can include various sub-modules (not shown in FIG. 1). Further, the memory 102 may include information pertaining to input(s)/output(s) of each step performed by the processor(s) 104 of the system 100 and methods of the present disclosure.
[029] The repository 102b may include a database or a data engine. Further, the repository 102b amongst other things, may serve as a database or includes a plurality of databases for storing the data that is processed, received, or generated as a result of the execution of the plurality of modules 102a. Although the repository 102b is shown internal to the system 100, it will be noted that, in alternate embodiments, the repository 102b can also be implemented external to the system 100, where the repository 102b may be stored within an external database (not shown in FIG. 1) communicatively coupled to the system 100. The data contained within such external database may be periodically updated. For example, data may be added into the external database and/or existing data may be modified and/or non-useful data may be deleted from the external database. In one example, the data may be stored in an external system, such as a Lightweight Directory Access Protocol (LDAP) directory and a Relational Database Management System (RDBMS). In another embodiment, the data stored in the repository 102b may be distributed between the system 100 and the external database.
[030] Referring to FIGS. 2A and 2B, components and functionalities of the system 100 are described in accordance with an example embodiment of the present disclosure. For example, FIGS. 2A and 2B illustrates exemplary flow diagrams of a processor-implemented method 200 for generating the cause-effect sentiment-enriched knowledge graph for portfolio optimization, in accordance with some embodiments of the present disclosure. Although steps of the method 200 including process steps, method steps, techniques or the like may be described in a sequential order, such processes, methods and techniques may be configured to work in alternate orders. In other words, any sequence or order of steps that may be described does not necessarily indicate a requirement that the steps be performed in that order. The steps of processes described herein may be performed in any practical order. Further, some steps may be performed simultaneously, or some steps may be performed alone or independently.
[031] At step 202 of the method 200, the one or more hardware processors 104 of the system 100 are configured to obtain one or more macroeconomic variables, one or more impacts for each of the one or more macroeconomic variables, and a change in a stock. In an embodiment, the one or more macroeconomic variables and the one or more impacts for each of the one or more macroeconomic variables, are obtained from one or more news articles, using a natural language processing (NLP) technique.
[032] Each of the one or more news articles comprises one or more financial statements and each of the one or more financial statements includes one or more sentences and further each sentence includes one or more words. In an embodiment, the one or more news articles may be released or presented by a government or a financial department of the certain government or may be provided by the relevant organization or entity having a financial or business relationship of the source organization for which the efficient knowledge graph to be made. The natural language processing (NLP) technique is used to extract the words that implies the financial entities present in each of the one or more news articles. The extraction of financial words using the NLP technique by parsing the sentence and tokenizing them then applying named entity recognition (NER) technique. The financial entities are recognized using the financial ontology which helps in identifying list of financial words.
[033] The one or more impacts for each of the one or more macroeconomic variables, are basically the impacting words for each macroeconomic variables or between two or more macroeconomic variables. The natural language processing (NLP) is used to extract such impacts (words) that implies the financial entities present in each of the one or more news articles. The extraction of impact words using the NLP technique by parsing the sentence and tokenizing them then applying named entity recognition (NER). The financial entities are recognized using the financial ontology which helps in identifying list of financial words.
[034] An exemplary news article is mentioned below to explain the macroeconomic variables and the impacts for each of the macroeconomic variables:
Exemplary news article: ‘Supply concerns are buoying the market, as Western sanctions on Russian oil bite. The push and pull between supply concerns and uncertainty over global growth in the face of inflation and rising interest rates are likely to play out in the market for some time, analysts said. He said it is unclear how big a risk there is of demand destruction, given the global economy is still recovering from the COVID slump.’
[035] The macroeconomic variables present in the above-mentioned exemplary news article, are: {‘Russian oil’, ‘global growth’, ‘demand and supply’}. The financial entities present in the above-mentioned exemplary news article, are: {‘inflation’, ‘interest rate’}.
[036] The change in the stock indicates the stock exchange information of a certain organization or from multiple organization that influences the financial and stock trend in a certain geography. In an embodiment, the change in the stock information is extracted from the stock return sources or stock indices using the NLP technique. In an embodiment, the stock information is either downloaded from known sources, or scraped using a custom build crawler. The custom build crawler helps in crawling the latest stock information and as well as historical and latest stock information from different stock exchange sources. Exemplary time-stamped stock indices information is mentioned below:
• ABC Ltd., 1,495.55, 387.15 (11.09%), 20,15,013.89 Cr
• XYZ Corporation Ltd.,43.67, 90.09 (21.42%), 1,96,307.17 Cr
• KAC Corporation Ltd.,85.17, -5.80 (-4.57%), 49,617.25 Cr
[037] At step 204 of the method 200, the one or more hardware processors 104 of the system 100 are configured to obtain a definition associated with each of the one or more macroeconomic variables, using financial ontologies and a sentiment dictionary. The financial ontologies comprise definitions of each financial word or financial entity or other words the macroeconomic variable. The sentiment dictionary contains the financial entity words or other words related to macroeconomics with their sentiment. Hence the definition and the sentiment associated with each of the one or more macroeconomic variables are extracted from the financial ontologies and a sentiment dictionary, using the NLP technique. The approach is to convert each word to its root form using lemmatizatizer and match it to identify similar words.
[038] The exemplary definition for some of the macroeconomic variables obtained from the exemplary news article mentioned at step 202 of the method 200 are provided below:
•
•
•
[039] At step 206 of the method 200, the one or more hardware processors 104 of the system 100 are configured to obtain a time-stamped financial trend information associated with each of the one or more macroeconomic variables, using one or more geography economy sources. The time-stamped financial trend information associated with each of the one or more macroeconomic variables, defines the historical and present trend of the market based on the continuous change in the stock information. The time-stamped financial trend information associated with each of the one or more macroeconomic variables, is obtained from one or more geography economy sources such as financial department of the certain government or a national bank of the country, and so on. The one or more geography economy sources may list those the financial trend information in the official websites or official announcement through press release, and so on.
[040] In an embodiment, the time-stamped financial information is obtained from company stock indices and some exemplary stock indices include:
• Timestamp1, ABC Ltd., 2,492.55, 397.75 (18.99%), 16,95,016.82 Cr
• Timestamp2, XY Corporation Ltd.,134.15, 21.8 (19.4%), 1,69,07.97 Cr
• Timestamp3, BCD Ltd,105.95, -3.90 (-3.55%), 99,413.52 Cr
[041] At step 208 of the method 200, the one or more hardware processors 104 of the system 100 are configured to determine one or more triplets from: (i) the one or more macroeconomic variables and (ii) the one or more impacts for each of the one or more macroeconomic variables, obtained at step 202 of the method 200. Again, the natural language processing (NLP) technique is used to determine the one or more triplets. As the name suggests the triplet is made of three components and in general it consists of a subject, an object and a relation between the subject and the object. The relation mostly be a verb. In the present disclosure, each triplet includes at least one of either: (i) a macroeconomic variable of the one or more macroeconomic variables, or (ii) an impact of the one or more impacts. Here the macroeconomic variable and the impact plays the roles of the subject and the object. However, the subject may sometimes replace with the object, and vice versa.
[042] Further understanding of sentence structure, additionally important identified financial entities are used, and subject-verb-object (SVO) are extracted. The financial entities and macroeconomic words act as a base to understand the important subjects in the documents. Further based on the sentence structure and built dependency tree, SVO patterns are identified and extracted. Any noise from the identified SVO patterns is removed. This helps in extracting the SVOs with proper financial domain key terms.
[043] In an embodiment, the triplet includes one macroeconomic variable, one impact and the relation between the macroeconomic variable and the impact. In another embodiment, the triplet includes two macroeconomic variables, and the relation between the two macroeconomic variables. In another embodiment, the triplet includes two impacts, and the relation between the two impacts.
[044] The exemplary triplets determined from the exemplary news article mentioned at step 202 of the method 200 are provided below:
•
•
•
•
•
•
•
[045] At step 210 of the method 200, the one or more hardware processors 104 of the system 100 are configured to build a base knowledge graph using the one or more triplets obtained at step 208 of the method 200. More specifically, the one or more macroeconomic variables, or the impacts or both present in each triplet are represented in the form of the base knowledge graph using the relation present in each triplet as well.
[046] According to an embodiment of the disclosure, the base knowledge graph includes a plurality of nodes and one or more edges wherein each edge is present between the two nodes. Each node of the plurality of nodes is represented by either a macroeconomic variable or an impact present in each of the one or more triplets. Each edge of the one or more edges is represented by the relation present in the corresponding triplet and thus defines the relation between the two nodes. In an embodiment, a graph database and the natural language processing (NLP) technique are used to build the base knowledge graph from the macroeconomic variables and the impacts present in the triplets.
[047] The base knowledge graph though it is built using the microeconomic variables and the impacts in the structured way, is not a complete knowledge graph. It is inefficient to use for deriving the effective financial insights, as much more relations and inter-relations between the macroeconomic variables and the impacts, such as numerical relations (such as increase, decrease) and the effect of the numerical relations (increase sometime positive or negative), sentimental information between the macroeconomic variables and the impacts, and so on, to be updated for the efficient knowledge graph.
[048] FIG. 3 shows an exemplary based knowledge graph build from the one or more triplets, in accordance with some embodiments of the present disclosure. The exemplary based knowledge graph of FIG. 3 is built for the exemplary news article mentioned at step 202 of the method 200. As shown in FIG. 3, the macroeconomic variable ‘Global Growth’ effects (impact) the financial entity ‘Interest Rates’ which belongs to the Bank. Similarly, the macroeconomic variable ‘Rise’ of (impact) the financial entity ‘Inflation’ which effects (impact) by the financial entity ‘Interest Rates’, and so on.
[049] At step 212 of the method 200, the one or more hardware processors 104 of the system 100 are configured to determine (i) one or more numerical relations associated with each triplet, (ii) one or more numerical sentiments associated with each triplet, (iii) one or more events and cooccurrence of the one or more events associated with each triplet, (iv) one or more company names and associated sustainability rankings associated with each triplet, and (v) cooccurrence of each microeconomic variable present in each triplet.
[050] The one or more numerical relations associated with each triplet is typically the numerical relations (for example, increase or decrease) for the financial entities (either microeconomic variable or the impact) present in the associated triplet. In an embodiment, the one or more numerical relations associated with each triplet are determined using the below five steps. At first step, the one or more financial entities present in the corresponding triplet are extracted. As detailed, the one or more financial entities are either the macroeconomic variable or the impact for the macroeconomic variable, or both as present in the corresponding triplet.
[051] At second step, one or more connecting verbs associated with the each of the one or more financial entities are annotated using the financial ontologies received at step 202 of the method 200. The financial ontologies are used to find the connecting verbs for each financial entity and are annotated to each financial entity. At the third step, one or more sentiments for each financial entity of the one or more financial entities, are determined using the NLP technique and the sentiment dictionary. The sentiment dictionary is used to identify the sentiment words for each financial entity through the NLP technique.
[052] At the fourth step, a pre-trained word vector model is used and refined with the extracted one or more financial entities at the first step, the annotations of the connected verbs associated with the extracted one or more financial entities determined at the second step, and the determined one or more sentiments for each financial entity of the one or more financial entities determined at the third step, to obtain the word vector model. The obtained vector model is a refined word vector model associated with each triplet.
[053] At the fifth and last step, the extracted one or more financial entities, the annotations of the connected verbs associated with the extracted one or more financial entities and the determined one or more sentiments for each financial entity of the one or more financial entities, to the word vector model refined at fourth step, to obtain the one or more numerical relations associated with each triplet.
[054] Next, the one or more numerical sentiments associated with each triplet are typically the numerical sentiments for the financial entities (either microeconomic variable or the impact) present in the associated triplet. In an embodiment, the one or more numerical sentiments associated with each triplet are determined using a trained neural network model.
[055] Exemplary numerical sentiments from the exemplary news article mentioned at step 202 of the method 200 are provided below:
• {Crude Oil International price (positive) -> decrease (negative) -> Crude Oil Import (positive)}
• {Inflation rate (negative)-> relates to (neutral) -> seasonal effect (positive)}
• {Money Supply (positive) -> Increase (negative) -> Inflation rate (negative)}
[056] In an embodiment, the trained neural network model is obtained using the below seven steps. At the first step, the one or more financial entities present in the corresponding triplet are extracted. As detailed, the one or more financial entities are either the macroeconomic variable or the impact for the macroeconomic variable, or both as present in the corresponding triplet. At the second step, one or more embeddings are obtained from the one or more financial entities present in the triplet, using a word vector model. Each embedding is associated with each financial entity.
[057] At the third step, one or more sentiments for each financial entity of the one or more financial entities, are determined using the NLP technique and the sentiment dictionary. The sentiment dictionary is used to identify the sentiment words for each financial entity through the NLP technique. At fourth step, the definition associated with each of the one or more financial entities present in the triplet, is obtained using the financial ontologies received at step 202 of the method 200.
[058] At the fifth step, a label for each triplet, is assigned based on the one or more financial entities present in the triplet. For example, if the financial entity present in the triplet is the microeconomic variable, then such triplet is labelled with ‘1’. Similarly, if the financial entity present in the triplet is not the microeconomic variable, then such triplet is labelled with ‘0’. At the sixth step, the one or more sentiments for each triplet determined at third step are assigned to each triple, based on the label assigned for the corresponding triplet at fifth step. For example, if the label assigned to the triplet is ‘1’, then the determined sentiment is assigned for such triplet as the positive sentiment. Similarly, if the label assigned to the triplet is ‘0’, then the determined sentiment is assigned for such triplet as the negative sentiment.
[059] At seventh step, a neural network model is trained with (i) the one or more embeddings from the one or more financial entities present in the triplet, at the second step (ii) the one or more sentiments associated with the triplet, determined at third step (iii) the definition associated with the one or more financial entities present in the triplet, obtained at fourth step (iv) the assigned label for each triplet at fifth step, and (v) the assigned one or more sentiments for each triplet at sixth step, to obtain the trained neural network model.
[060] The trained neural network model obtained at the seventh step is then used to determine the one or more numerical sentiments associated with each triplet, by passing the one or more financial entities present in the triplet to the trained neural network model.
[061] Exemplary triplets having the sentimental relation are mentioned below, wherein the words like ‘neutral’, ‘negative’, ‘positive’ are the sentiments:
•
•
•
•
•
•
•
[062] In an embodiment, the one or more events (for example, contagious disease such as COVID, war between the countries) and the cooccurrence of the one or more events associated with each triplet, are determined from (i) the one or more macroeconomic variables, and (ii) the one or more impacts for each of the one or more macroeconomic variables, using the NLP technique. For this, if the triplet contains the macroeconomic variable, then the impacts for such macroeconomic variable are identified. Then finally, the macroeconomic variable and the impacts for such macroeconomic variable are used to identify the the one or more events and the cooccurrence of the one or more events associated with the corresponding triplet.
[063] Exemplary events are listed below:
•
•
•
[064] In an embodiment, the one or more company names and the associated sustainability rankings associated with each triplet are determined, from annual reports, using the NLP technique.
[065] Table 1 shows exemplary sustainability rankings and the risk level for the exemplary companies:
Company ESG Risk Rating Risk
ABC Oil Corporation Ltd.
74.1 Severe Risk
XYZ Industries Ltd. 54.9 Severe Risk
IJK Corporation Ltd.
15.6 High Risk
Table 1
[066] In an embodiment, the cooccurrence of each microeconomic variable present in each triplet is determined, from the cooccurrence of the one or more events associated with the one or more macroeconomic variables, using the NLP technique. Exemplary cooccurrences from the exemplary news article mentioned at step 202 of the method 200 is:
•
•
[067] At step 214 of the method 200, the one or more hardware processors 104 of the system 100 are configured to update the base knowledge graph obtained at step 210 of the method 200 with (i) the one or more numerical relations associated with each triplet, (ii) the one or more numerical sentiments associated with each triplet, (iii) the one or more events and cooccurrence of the one or more events associated with each triplet, (iv) one or more company names and associated sustainability rankings associated with each triplet, (v) the cooccurrence of each microeconomic variable present in each triplet, and (vi) the change in the stock, to obtain the cause-effect sentiment-enriched knowledge graph. The cause-effect sentiment-enriched knowledge graph is complete and efficient having the complete information including definitions between the macroeconomic variables, numerical relations and the sentiments associated with the macroeconomic variables, the events and the cooccurrence of the one or more events associated with the macroeconomic variables.
[068] FIG. 4 shows an exemplary cause-effect sentiment-enriched knowledge graph for portfolio optimization, in accordance with some embodiments of the present disclosure. The exemplary cause-effect sentiment-enriched knowledge graph of FIG. 4 is built for the exemplary news article mentioned at step 202 of the method 200. As shown in FIG. 4, the macroeconomic variable ‘Rise’ of (impact) the financial entity ‘Inflation’ which effects (impact) by the financial entity ‘Interest Rates’, is a negative sentiment. Similarly, the event ‘Russian oil’ is a positive sentiment for the financial entity ‘crude oil’. Similarly, the event of ‘COVID 19’ effect the financial entity ‘economy’ which is a negative sentiment and so on.
[069] The generated cause-effect sentiment-enriched knowledge graph is continuously updated from time to time, with the business news, the events, the sentiment effects the events happening around the globe with the use of the domain knowledge. The new macroeconomic variables and the new impacts associated with the new macroeconomic variables are added to the cause-effect sentiment-enriched knowledge graph from time to time.
[070] At step 216 of the method 200, the one or more hardware processors 104 of the system 100 are configured to perform a reasoning on the cause-effect sentiment-enriched knowledge graph obtained at step 214 of the method 200, to extract financial insights and patterns for the portfolio management optimization. A rule engine having the set of rules associated with the required financial insights are passed to the cause-effect sentiment-enriched knowledge graph, to extract the financial insights and the patterns. The extracted financial insights and the patterns are used for various financial operations and advisory such as investments, stock sales, returns, and so on.
[071] The method and systems of the present disclosure provides the cause-effect sentiment-enriched knowledge graph by combining the domain knowledge such as financial ontologies, microeconomic factors, and the business knowledge in picture considering the events happening around the words, sentimental relations, numerical relations associated with the sentimental factors and so on, along with the news across the globe. Further, the cause-effect sentiment-enriched knowledge graph is continuously updated from time to time, with the business news, the events, the sentiment effects the events happening around the globe with the use of the domain knowledge. The new macroeconomic variables and the new impacts associated with the new macroeconomic variables are added to the cause-effect sentiment-enriched knowledge graph from time to time. Hence the generated the cause-effect sentiment-enriched knowledge graph is complete, accurate and efficient for portfolio optimization.
[072] The generated cause-effect sentiment-enriched knowledge graph is efficient for analyzing and predicting the market trends with their macroeconomic factors and their movement. The best optimized results are achieved from the obtained cause-effect sentiment-enriched knowledge graph, as the news information is also included to understand the market with the objective of effective portfolio optimization Further, the cause-effect sentiment-enriched knowledge graph can also be used to control the financial risks in the business activities.
[073] The embodiments of present disclosure herein address unresolved problem of generating the efficient and complete cause-effect sentiment-enriched knowledge graph for analyzing and predicting the market trends with their macroeconomic factors and their movement. The generated cause-effect sentiment-enriched knowledge graph is not only used for seeking the right guidance from various financial strategies for the given level of risk, but also controlling the financial risks and at the same time to yield most possible returns.
[074] The written description describes the subject matter herein to enable any person skilled in the art to make and use the embodiments. The scope of the subject matter embodiments is defined by the claims and may include other modifications that occur to those skilled in the art. Such other modifications are intended to be within the scope of the claims if they have similar elements that do not differ from the literal language of the claims or if they include equivalent elements with insubstantial differences from the literal language of the claims.
[075] It is to be understood that the scope of the protection is extended to such a program and in addition to a computer-readable means having a message therein; such computer-readable storage means contain program-code means for implementation of one or more steps of the method, when the program runs on a server or mobile device or any suitable programmable device. The hardware device can be any kind of device which can be programmed including e.g., any kind of computer like a server or a personal computer, or the like, or any combination thereof. The device may also include means which could be e.g., hardware means like e.g., an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a combination of hardware and software means, e.g., an ASIC and an FPGA, or at least one microprocessor and at least one memory with software processing components located therein. Thus, the means can include both hardware means and software means. The method embodiments described herein could be implemented in hardware and software. The device may also include software means. Alternatively, the embodiments may be implemented on different hardware devices, e.g., using a plurality of CPUs.
[076] The embodiments herein can comprise hardware and software elements. The embodiments that are implemented in software include but are not limited to, firmware, resident software, microcode, etc. The functions performed by various components described herein may be implemented in other components or combinations of other components. For the purposes of this description, a computer-usable or computer readable medium can be any apparatus that can comprise, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device.
[077] The illustrated steps are set out to explain the exemplary embodiments shown, and it should be anticipated that ongoing technological development will change the manner in which particular functions are performed. These examples are presented herein for purposes of illustration, and not limitation. Further, the boundaries of the functional building blocks have been arbitrarily defined herein for the convenience of the description. Alternative boundaries can be defined so long as the specified functions and relationships thereof are appropriately performed. Alternatives (including equivalents, extensions, variations, deviations, etc., of those described herein) will be apparent to persons skilled in the relevant art(s) based on the teachings contained herein. Such alternatives fall within the scope of the disclosed embodiments. Also, the words “comprising,” “having,” “containing,” and “including,” and other similar forms are intended to be equivalent in meaning and be open ended in that an item or items following any one of these words is not meant to be an exhaustive listing of such item or items, or meant to be limited to only the listed item or items. It must also be noted that as used herein and in the appended claims, the singular forms “a,” “an,” and “the” include plural references unless the context clearly dictates otherwise.
[078] Furthermore, one or more computer-readable storage media may be utilized in implementing embodiments consistent with the present disclosure. A computer-readable storage medium refers to any type of physical memory on which information or data readable by a processor may be stored. Thus, a computer-readable storage medium may store instructions for execution by one or more processors, including instructions for causing the processor(s) to perform steps or stages consistent with the embodiments described herein. The term “computer-readable medium” should be understood to include tangible items and exclude carrier waves and transient signals, i.e., be non-transitory. Examples include random access memory (RAM), read-only memory (ROM), volatile memory, nonvolatile memory, hard drives, CD ROMs, DVDs, flash drives, disks, and any other known physical storage media.
[079] It is intended that the disclosure and examples be considered as exemplary only, with a true scope of disclosed embodiments being indicated by the following claims.
, Claims:We Claim:
1. A processor-implemented method (200) for generating a cause-effect sentiment-enriched knowledge graph for portfolio optimization, comprising the steps of:
obtaining, via one or more hardware processors, one or more macroeconomic variables, one or more impacts for each of the one or more macroeconomic variables, and a change in a stock (202);
obtaining, via the one or more hardware processors, a definition associated with each of the one or more macroeconomic variables, using financial ontologies and a sentiment dictionary (204);
obtaining, via the one or more hardware processors, a time-stamped financial trend information associated with each of the one or more macroeconomic variables, using one or more geography economy sources (206);
determining, via the one or more hardware processors, one or more triplets from: (i) the one or more macroeconomic variables and (ii) the one or more impacts for each of the one or more macroeconomic variables, using a natural language processing (NLP) technique, wherein each triplet of the one or more triplets comprises at least one of either: (i) a macroeconomic variable of the one or more macroeconomic variables, or (ii) an impact of the one or more impacts (208);
building, via the one or more hardware processors, a base knowledge graph comprising a plurality of nodes and one or more edges, based on the one or more triplets, using a graph database and the natural language processing (NLP) technique, wherein each node of the plurality of nodes is represented by either the macroeconomic variable or the impact present in each of the one or more triplets, and each edge of the one or more edges is represented by a relation between the two nodes of the plurality of nodes (210);
determining, via the one or more hardware processors, (i) one or more numerical relations associated with each triplet, (ii) one or more numerical sentiments associated with each triplet, (iii) one or more events and cooccurrence of the one or more events associated with each triplet, (iv) one or more company names and associated sustainability rankings associated with each triplet, and (v) cooccurrence of each microeconomic variable present in each triplet (212); and
updating, via the one or more hardware processors, the base knowledge graph with (i) the one or more numerical relations associated with each triplet, (ii) the one or more numerical sentiments associated with each triplet, (iii) the one or more events and cooccurrence of the one or more events associated with each triplet, (iv) one or more company names and associated sustainability rankings associated with each triplet, (v) the cooccurrence of each microeconomic variable present in each triplet, and (vi) the change in the stock, to obtain the cause-effect sentiment-enriched knowledge graph (214).

2. The method of claim 1, further comprising: performing, via the one or more hardware processors, a reasoning on the cause-effect sentiment-enriched knowledge graph, using a rule engine, to extract financial insights and patterns for the portfolio management optimization (216).

3. The method of claim 1, wherein:
the one or more macroeconomic variables and the one or more impacts for each of the one or more macroeconomic variables, are obtained from one or more news articles, using the natural language processing (NLP) technique, wherein each of the one or more news articles comprises one or more financial statements; and
the change in the stock is obtained from stock return sources, using the natural language processing (NLP) technique.

4. The method of claim 1, wherein the one or more numerical relations associated with each triplet, are determined by:
extracting one or more financial entities present in the triplet, wherein the one or more financial entities are either the macroeconomic variable or the impact for the macroeconomic variable;
annotating one or more connecting verbs associated with the one or more financial entities, using the financial ontologies;
determining one or more sentiments for each financial entity of the one or more financial entities, using the NLP technique and the sentiment dictionary;
refining a pre-trained word vector model, with the extracted one or more financial entities, the annotations of the connected verbs associated with the extracted one or more financial entities and the determined one or more sentiments for each financial entity of the one or more financial entities, to obtain the word vector model; and
passing the extracted one or more financial entities, the annotations of the connected verbs associated with the extracted one or more financial entities and the determined one or more sentiments for each financial entity of the one or more financial entities, to the word vector model, to obtain the one or more numerical relations associated with each triplet.

5. The method of claim 1, wherein the one or more numerical sentiments associated with each triplet, are determined using a trained neural network model, and wherein the trained neural network model is obtained by:
extracting one or more financial entities present in the triplet, wherein the one or more financial entities are either the macroeconomic variable or the impact for the macroeconomic variable;
obtaining one or more embeddings from the one or more financial entities present in the triplet, using a word vector model;
determining one or more sentiments for each financial entity of the one or more financial entities, using the NLP technique and the sentiment dictionary;
obtaining the definition associated with each of the one or more financial entities present in the triplet, using the financial ontologies;
assigning a label for each triplet, based on the one or more financial entities present in the triplet;
assigning the determined one or more sentiments for each triplet, based on the label assigned for each triplet; and
training a neural network model with (i) the one or more embeddings from the one or more financial entities present in the triplet, (ii) the one or more sentiments associated with the triplet, (iii) the definition associated with the one or more financial entities present in the triplet, (iv) the assigned label for each triplet, and (v) the assigned one or more sentiments for each triplet, to obtain the trained neural network model.

6. The method of claim 1, wherein the one or more events and the cooccurrence of the one or more events associated with each triplet, are determined from (i) the one or more macroeconomic variables, and (ii) the one or more impacts for each of the one or more macroeconomic variables, using the NLP technique.

7. The method of claim 1, wherein the one or more company names and the associated sustainability rankings associated with each triplet are determined, from annual reports, using the NLP technique.

8. The method of claim 1, wherein the cooccurrence of each microeconomic variable present in each triplet are determined, from the cooccurrence of the one or more events associated with the one or more macroeconomic variables, using the NLP technique.

9. A system (100) for generating a cause-effect sentiment-enriched knowledge graph for portfolio optimization, comprising:
a memory (102) storing instructions;
one or more input/output (I/O) interfaces (106); and
one or more hardware processors (104) coupled to the memory (102) via the one or more I/O interfaces (106), wherein the one or more hardware processors (104) are configured by the instructions to:
obtain one or more macroeconomic variables, one or more impacts for each of the one or more macroeconomic variables, and a change in a stock;
obtain a definition associated with each of the one or more macroeconomic variables, using financial ontologies and a sentiment dictionary;
obtain a time-stamped financial trend information associated with each of the one or more macroeconomic variables, using one or more geography economy sources;
determine one or more triplets from: (i) the one or more macroeconomic variables and (ii) the one or more impacts for each of the one or more macroeconomic variables, using a natural language processing (NLP) technique, wherein each triplet of the one or more triplets comprises at least one of either: (i) a macroeconomic variable of the one or more macroeconomic variables, or (ii) an impact of the one or more impacts;
build a base knowledge graph comprising a plurality of nodes and one or more edges, based on the one or more triplets, using a graph database and the natural language processing (NLP) technique, wherein each node of the plurality of nodes is represented by either the macroeconomic variable or the impact present in each of the one or more triplets, and each edge of the one or more edges is represented by a relation between the two nodes of the plurality of nodes;
determine (i) one or more numerical relations associated with each triplet, (ii) one or more numerical sentiments associated with each triplet, (iii) one or more events and cooccurrence of the one or more events associated with each triplet, (iv) one or more company names and associated sustainability rankings associated with each triplet, and (v) cooccurrence of each microeconomic variable present in each triplet; and
update the base knowledge graph with (i) the one or more numerical relations associated with each triplet, (ii) the one or more numerical sentiments associated with each triplet, (iii) the one or more events and cooccurrence of the one or more events associated with each triplet, (iv) one or more company names and associated sustainability rankings associated with each triplet, (v) the cooccurrence of each microeconomic variable present in each triplet, and (vi) the change in the stock, to obtain the cause-effect sentiment-enriched knowledge graph.

10. The system of claim 9, wherein the one or more hardware processors (104) are configured to perform a reasoning on the cause-effect sentiment-enriched knowledge graph, using a rule engine, to extract financial insights and patterns for the portfolio management optimization.

11. The system of claim 9, wherein the one or more hardware processors (104) are configured obtain:
the one or more macroeconomic variables and the one or more impacts for each of the one or more macroeconomic variables, from one or more news articles, using the natural language processing (NLP) technique, wherein each of the one or more news articles comprises one or more financial statements; and
the change in the stock, from stock return sources, using the natural language processing (NLP) technique.

12. The system of claim 9, wherein the one or more hardware processors (104) are configured to determine the one or more numerical relations associated with each triplet, by:
extracting one or more financial entities present in the triplet, wherein the one or more financial entities are either the macroeconomic variable or the impact for the macroeconomic variable;
annotating one or more connecting verbs associated with the one or more financial entities, using the financial ontologies;
determining one or more sentiments for each financial entity of the one or more financial entities, using the NLP technique and the sentiment dictionary;
refining a pre-trained word vector model, with the extracted one or more financial entities, the annotations of the connected verbs associated with the extracted one or more financial entities and the determined one or more sentiments for each financial entity of the one or more financial entities, to obtain the word vector model; and
passing the extracted one or more financial entities, the annotations of the connected verbs associated with the extracted one or more financial entities and the determined one or more sentiments for each financial entity of the one or more financial entities, to the word vector model, to obtain the one or more numerical relations associated with each triplet.

13. The system of claim 9, wherein the one or more hardware processors (104) are configured to determine the one or more numerical sentiments associated with each triplet, using a trained neural network model, wherein the trained neural network model is obtained by:
extracting one or more financial entities present in the triplet, wherein the one or more financial entities are either the macroeconomic variable or the impact for the macroeconomic variable;
obtaining one or more embeddings from the one or more financial entities present in the triplet, using a word vector model;
determining one or more sentiments for each financial entity of the one or more financial entities, using the NLP technique and the sentiment dictionary;
obtaining the definition associated with each of the one or more financial entities present in the triplet, using the financial ontologies;
assigning a label for each triplet, based on the one or more financial entities present in the triplet;
assigning the determined one or more sentiments for each triplet, based on the label assigned for each triplet; and
training a neural network model with (i) the one or more embeddings from the one or more financial entities present in the triplet, (ii) the one or more sentiments associated with the triplet, (iii) the definition associated with the one or more financial entities present in the triplet, (iv) the assigned label for each triplet, and (v) the assigned one or more sentiments for each triplet, to obtain the trained neural network model.

14. The system of claim 9, wherein the one or more hardware processors (104) are configured to determine the one or more events and the cooccurrence of the one or more events associated with each triplet, from (i) the one or more macroeconomic variables, and (ii) the one or more impacts for each of the one or more macroeconomic variables, using the NLP technique.

15. The system of claim 9, wherein the one or more hardware processors (104) are configured to determine the one or more company names and the associated sustainability rankings associated with each triplet, from annual reports, using the NLP technique.

16. The system of claim 9, wherein the one or more hardware processors (104) are configured to determine the cooccurrence of each microeconomic variable present in each triplet, from the cooccurrence of the one or more events associated with the one or more macroeconomic variables, using the NLP technique.

Dated this 30th Day of November 2022

Tata Consultancy Services Limited
By their Agent & Attorney

(Adheesh Nargolkar)
of Khaitan & Co
Reg No IN-PA-1086