DESC:
FORM 2
THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENTS RULES, 2003

COMPLETE SPECIFICATION
(See section 10 and rule 13)
1. TITLE OF THE INVENTION
METHOD AND SYSTEM FOR RECOMMENDING MODELS FOR FORECASTING
2. APPLICANT(S)
NAME NATIONALITY ADDRESS
JIO PLATFORMS LIMITED INDIAN OFFICE-101, SAFFRON, NR. CENTRE POINT, PANCHWATI 5 RASTA, AMBAWADI, AHMEDABAD 380006, GUJARAT, INDIA
3.PREAMBLE TO THE DESCRIPTION

THE FOLLOWING SPECIFICATION PARTICULARLY DESCRIBES THE NATURE OF THIS INVENTION AND THE MANNER IN WHICH IT IS TO BE PERFORMED.

FIELD OF THE INVENTION
[0001] The present invention relates to forecasting service, more particularly relates to a method and a system for recommending models for forecasting.
BACKGROUND OF THE INVENTION
[0002] In general, the communication network is monitored by monitoring all the core networking components. The intent of monitoring these networking components is to detect faults or anomalies, if any.
[0003] Apart from just monitoring the communication network, various forecasting models are being utilized in order to forecast occurrence of future events such as faults and anomalies in the communication network. However, in order to forecast occurrence of any anomalies or faults in the communication network, generally pre-set forecasting models/logics are being used. Due to the nature of ever evolving communication networks, new types of anomalies or issues are being detected every now and then. Therefore, usage of pre-set forecasting logics for new types of anomalies or issues is not the appropriate approach. In other words, the accuracy of forecasting by the preset forecasting logics may be low. Further, due to less accuracy due to usage of present pre-set forecasting logics, forecasting of future events makes it difficult to identify potential issues in advance, which may eventually disrupt the communication network.
[0004] In view of the above, there is a dire need for an efficient system and method for forecasting future events, which ensures better accuracy of forecasting of future events and provides enhanced consumer experience.
SUMMARY OF THE INVENTION
[0005] One or more embodiments of the present disclosure provide a method and system for recommending models for forecasting.
[0006] In one aspect of the present invention, the system for recommending the models for forecasting is disclosed. The system includes a retrieving unit configured to retrieve data from one or more data sources. The system further includes an extracting unit, configured to extract one or more features from the retrieved and pre-processed data. The system further includes a configuring unit, to configure, one or more hyperparameters for at least one of a plurality of models which are required to be trained. The system further includes a selecting unit, configured to select, a training date range and a forecasting date range for the at least one of the plurality of models. The system further includes a training unit, configured to, train, the at least one of the plurality of models with at least one of, the retrieved data, the extracted one or more features, the configured one or more hyperparameters and the selected training date range. The system further includes a forecasting engine, configured to forecast one or more events using at least one of the plurality of trained models based on the selected forecasting date range. The system further includes a recommending unit, configured to recommend, one or more trained models for forecasting using results generated by the at least one of the trained models.
[0007] In an embodiment, the one or more data sources include at least one of, Hadoop Distributed File System (HDFS), Non-Access Stratum (NAS), streaming engine, Hypertext Transfer Protocols (HTTP2).
[0008] In an embodiment, the retrieving unit is further configured to perform data definition on the retrieved data. Upon performing the data definition, the retrieving unit is configured to pre-process the retrieved data and store the pre-processed data in a storage unit.
[0009] In an embodiment, the one or more features are extracted from the retrieved data depending on the type of training required to be performed.
[0010] In an embodiment, the selecting unit, selects, the training date range for the at least one of the plurality of models by at least one of setting, a threshold of a number of datasets to be provided to the at least one of the plurality of models for training or setting, a time range to the at least one of the plurality of models for training.
[0011] In an embodiment, the training date range is at least one of, dynamically set by the selecting unit depending on the type of data training or, the training date range is set by the selecting unit based on receiving an input from a user.
[0012] In an embodiment, the one or more hyperparameters are automatically configured by the configuring unit, or the one or more hyperparameters are configured by the user.
[0013] In an embodiment, the at least one of the plurality of models learns patterns/trends based on the retrieved data.
[0014] In an embodiment, the recommending unit, recommends, one or more trained models for forecasting using results generated by each of the trained models, by generating, the results of the at least one of the plurality of trained models, wherein the results includ, but not limited to, accuracy values of forecasting by the at least one of the plurality of trained models, Root Mean Square Error (RMSE) values of the at least one of the plurality of trained models, training status list for the at least one of the plurality of models and forecasting status list for the at least one of the plurality of trained models. Further, the recommending unit recommends the one or more trained models based on the generated results including at least one of, the accuracy values of forecasting by the at least one of the plurality of trained models and the RMSE values of the at least one of the plurality of trained models.
[0015] In another aspect of the present invention, the method for recommending the models for forecasting is disclosed. The method includes the step of retrieving data from one or more data sources. The method further includes the step of extracting one or more features from the retrieved and pre-processed data. The method further includes the step of configuring one or more hyperparameters for at least one of a plurality of models which are required to be trained. The method further includes the step of selecting a training date range and forecasting date range for the at least one of the plurality of models. The method further includes training the at least one of the plurality of models with at least one of, the retrieved data, the extracted one or more features, the configured one or more hyperparameters and the selected training date range. The method further includes forecasting one or more events using the at least one of the plurality of trained models based on the selected training date range. The method further includes recommending one or more trained models for forecasting using results generated by each of the trained models.
[0016] In another aspect of the invention, a non-transitory computer-readable medium having stored thereon computer-readable instructions is disclosed. The computer-readable instructions are executed by a processor. The processor is configured to retrieve data from one or more data sources. The processor is configured to extract one or more features from the retrieved and pre-processed data. The processor is configured to configure one or more hyperparameters for at least one of a plurality of models which are required to be trained. The processor is configured to select a training date range and a forecasting date range for the at least one of the plurality of models. The processor is configured to train, the at least one of the plurality of models with at least one of, the retrieved data, the extracted one or more features, the configured one or more hyperparameters and the selected training date range. The processor is configured to forecast one or more events using the at least one of the plurality of trained models based on the selected forecasting date range. The processor is configured to recommend one or more trained models for forecasting using results generated by each of the trained models.
[0017] Other features and aspects of this invention will be apparent from the following description and the accompanying drawings. The features and advantages described in this summary and in the following detailed description are not all-inclusive, and particularly, many additional features and advantages will be apparent to one of ordinary skill in the relevant art, in view of the drawings, specification, and claims hereof. Moreover, it should be noted that the language used in the specification has been principally selected for readability and instructional purposes and may not have been selected to delineate or circumscribe the inventive subject matter, resort to the claims being necessary to determine such inventive subject matter.
BRIEF DESCRIPTION OF THE DRAWINGS
[0018] The accompanying drawings, which are incorporated herein, and constitute a part of this disclosure, illustrate exemplary embodiments of the disclosed methods and systems in which like reference numerals refer to the same parts throughout the different drawings. Components in the drawings are not necessarily to scale, emphasis instead being placed upon clearly illustrating the principles of the present disclosure. Some drawings may indicate the components using block diagrams and may not represent the internal circuitry of each component. It will be appreciated by those skilled in the art that disclosure of such drawings includes disclosure of electrical components, electronic components or circuitry commonly used to implement such components.
[0019] FIG. 1 is an exemplary block diagram of an environment for recommending models for forecasting, according to one or more embodiments of the present invention;
[0020] FIG. 2 is an exemplary block diagram of a system for recommending the models for forecasting, according to one or more embodiments of the present invention;
[0021] FIG. 3 is an exemplary block diagram of an architecture implemented in the system of the FIG. 2, according to one or more embodiments of the present invention;
[0022] FIG. 4 is a flow diagram for recommending the models for forecasting, according to one or more embodiments of the present invention; and
[0023] FIG. 5 is a schematic representation of a method for recommending the models for forecasting, according to one or more embodiments of the present invention.
[0024] The foregoing shall be more apparent from the following detailed description of the invention.
DETAILED DESCRIPTION OF THE INVENTION
[0025] Some embodiments of the present disclosure, illustrating all its features, will now be discussed in detail. 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.
[0026] Various modifications to the embodiment will be readily apparent to those skilled in the art and the generic principles herein may be applied to other embodiments. However, one of ordinary skill in the art will readily recognize that the present disclosure including the definitions listed here below are not intended to be limited to the embodiments illustrated but is to be accorded the widest scope consistent with the principles and features described herein.
[0027] A person of ordinary skill in the art will readily ascertain that the illustrated steps detailed in the figures and here below 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 and spirit of the disclosed embodiments.
[0028] The present invention provides a unique approach of recommending the consumer with the appropriate forecasting models from the multiple forecasting models which is specific to the type of application the consumer intends to monitor. Thereafter, the consumer may select the recommended appropriate forecasting model from the multiple forecasting models in order to forecast future events.
[0029] FIG. 1 illustrates an exemplary block diagram of an environment 100 for recommending models for forecasting, according to one or more embodiments of the present disclosure. In this regard, the environment 100 includes a User Equipment (UE) 102, a server 104, a network 106 and a system 108 communicably coupled to each other for recommending the models for forecasting. In an embodiment, the forecasting refers to the process of using trained models to predict or estimate one or more future events or data values based on historical data and extracted features.
[0030] As per the illustrated embodiment and for the purpose of description and illustration, the UE 102 includes, but not limited to, a first UE 102a, a second UE 102b, and a third UE 102c, and should nowhere be construed as limiting the scope of the present disclosure. In alternate embodiments, the UE 102 may include a plurality of UEs as per the requirement. For ease of reference, each of the first UE 102a, the second UE 102b, and the third UE 102c, will hereinafter be collectively and individually referred to as the “User Equipment (UE) 102”.
[0031] In an embodiment, the UE 102 is one of, but not limited to, any electrical, electronic, electro-mechanical or an equipment and a combination of one or more of the above devices such as a smartphone, virtual reality (VR) devices, augmented reality (AR) devices, laptop, a general-purpose computer, desktop, personal digital assistant, tablet computer, mainframe computer, or any other computing device.
[0032] The environment 100 includes the server 104 accessible via the network 106. The server 104 may include, by way of example but not limitation, one or more of a standalone server, a server blade, a server rack, a bank of servers, a server farm, hardware supporting a part of a cloud service or system, a home server, hardware running a virtualized server, one or more processors executing code to function as a server, one or more machines performing server-side functionality as described herein, at least a portion of any of the above, some combination thereof. In an embodiment, the entity may include, but is not limited to, a vendor, a network operator, a company, an organization, a university, a lab facility, a business enterprise side, a defense facility side, or any other facility that provides service.
[0033] The network 106 includes, by way of example but not limitation, one or more of a wireless network, a wired network, an internet, an intranet, a public network, a private network, a packet-switched network, a circuit-switched network, an ad hoc network, an infrastructure network, a Public-Switched Telephone Network (PSTN), a cable network, a cellular network, a satellite network, a fiber optic network, or some combination thereof. The network 106 may include, but is not limited to, a Third Generation (3G), a Fourth Generation (4G), a Fifth Generation (5G), a Sixth Generation (6G), a New Radio (NR), a Narrow Band Internet of Things (NB-IoT), an Open Radio Access Network (O-RAN), and the like.
[0034] The network 106 may also include, by way of example but not limitation, at least a portion of one or more networks having one or more nodes that transmit, receive, forward, generate, buffer, store, route, switch, process, or a combination thereof, etc. one or more messages, packets, signals, waves, voltage or current levels, some combination thereof, or so forth. The network 106 may also include, by way of example but not limitation, one or more of a wireless network, a wired network, an internet, an intranet, a public network, a private network, a packet-switched network, a circuit-switched network, an ad hoc network, an infrastructure network, a Public-Switched Telephone Network (PSTN), a cable network, a cellular network, a satellite network, a fiber optic network, a VOIP or some combination thereof.
[0035] The environment 100 further includes the system 108 communicably coupled to the server 104 and the UE 102 via the network 106. The system 108 is configured to recommend the models for forecasting. As per one or more embodiments, the system 108 is adapted to be embedded within the server 104 or embedded as an individual entity.
[0036] Operational and construction features of the system 108 will be explained in detail with respect to the following figures.
[0037] FIG. 2 is an exemplary block diagram of the system 108 for recommending the models for forecasting, according to one or more embodiments of the present invention.
[0038] As per the illustrated embodiment, the system 108 includes one or more processors 202, a memory 204, a user interface 206, and a database 208. For the purpose of description and explanation, the description will be explained with respect to one processor 202 and should nowhere be construed as limiting the scope of the present disclosure. In alternate embodiments, the system 108 may include more than one processor 202 as per the requirement of the network 106. The one or more processors 202, hereinafter referred to as the processor 202 may be implemented as one or more microprocessors, microcomputers, microcontrollers, digital signal processors, central processing units, state machines, logic circuitries, single board computers, and/or any devices that manipulate signals based on operational instructions.
[0039] As per the illustrated embodiment, the processor 202 is configured to fetch and execute computer-readable instructions stored in the memory 204. The memory 204 may be configured to store one or more computer-readable instructions or routines in a non-transitory computer-readable storage medium, which may be fetched and executed to create or share data packets over a network service. The memory 204 may include any non-transitory storage device including, for example, volatile memory such as RAM, or non-volatile memory such as disk memory, EPROMs, FLASH memory, unalterable memory, and the like.
[0040] In an embodiment, the user interface 206 includes a variety of interfaces, for example, interfaces for a graphical user interface, a web user interface, a Command Line Interface (CLI), and the like. The user interface 206 facilitates communication of the system 108. In one embodiment, the user interface 206 provides a communication pathway for one or more components of the system 108. Examples of such components include, but are not limited to, the UE 102 and the database 208.
[0041] The database 208 is one of, but not limited to, a centralized database, a cloud-based database, a commercial database, an open-source database, a distributed database, an end-user database, a graphical database, a No-Structured Query Language (NoSQL) database, an object-oriented database, a personal database, an in-memory database, a document-based database, a time series database, a wide column database, a key value database, a search database, a cache databases, and so forth. The foregoing examples of database 208 types are non-limiting and may not be mutually exclusive e.g., a database can be both commercial and cloud-based, or both relational and open-source, etc.
[0042] In order for the system 108 to recommend the models for forecasting, the processor 202 includes one or more modules. In one embodiment, the one or more modules includes, but not limited to, a retrieving unit 210, a storage unit 212, an extracting unit 214, a configuring unit 216, a selecting unit 218, a training unit 220, a forecasting engine 222 and a recommending unit 224 communicably coupled to each other for recommending the models for forecasting.
[0043] In one embodiment, each of the one or more modules, the retrieving unit 210, the storage unit 212, the extracting unit 214, the configuring unit 216, the selecting unit 218, the training unit 220, the forecasting engine 222 and the recommending unit 224 can be used in combination or interchangeably for recommending the models for forecasting.
[0044] The retrieving unit 210, the storage unit 212, the extracting unit 214, the configuring unit 216, the selecting unit 218, the training unit 220, the forecasting engine 222 and the recommending unit 224 in an embodiment, may be implemented as a combination of hardware and programming (for example, programmable instructions) to implement one or more functionalities of the processor 202. In the examples described herein, such combinations of hardware and programming may be implemented in several different ways. For example, the programming for the processor 202 may be processor-executable instructions stored on a non-transitory machine-readable storage medium and the hardware for the processor may comprise a processing resource (for example, one or more processors), to execute such instructions. In the present examples, the memory 204 may store instructions that, when executed by the processing resource, implement the processor. In such examples, the system 108 may comprise the memory 204 storing the instructions and the processing resource to execute the instructions, or the memory 204 may be separate but accessible to the system 108 and the processing resource. In other examples, the processor 202 may be implemented by electronic circuitry.
[0045] In one embodiment, the retrieving unit 210 is configured to retrieve data from one or more data sources 302 (as shown in FIG. 3). The data refers to the information retrieved from the one or more sources that serves as input for the model training and forecasting processes. The one or more data sources 302 include, but not limited to, Hadoop Distributed File System (HDFS), Non-Access Stratum (NAS), streaming engine, Hypertext Transfer Protocols (HTTP2). The HDFS is a distributed storage system for storing large volumes of structured or unstructured data. The NAS data is related to signaling and communication in mobile networks, typically pertaining to higher-level network protocols and user sessions. The streaming engine refers to real-time data sources that provide continuous streams of data, such as log streams. The HTTP2 is a web-based data source that could involve fetching data from Application Programming Interfaces (APIs) or web servers.
[0046] Further, the retrieving unit 210 is configured to perform data definition on the retrieved data. The data definition refers to the process of specifying and structuring the retrieved data before it is further processed. The data definition includes identifying and categorizing the data, determining its attributes, establishing its format, and organizing it in a way that facilitates subsequent operations like feature extraction, model training, and storage. For example, where the retrieved data comes from an e-commerce platform's sales database, which contains information about product sales, customer details, and transactions. The retrieving unit 210 identifies the types of data retrieved, such as transaction data (order ID, order date, product ID), customer data (customer ID, location, age), product data (product name, category, price) etc. Upon identifying the data, the retrieving unit 210 categorizes the date into logical groups such as sales transactions, customer information, production information etc. For each category, the data definition process specifies the key attributes such as transaction ID, order data, product ID, quantity sold etc., under sales transaction, customer ID, age, location under customer information, product ID, product name, price, category under product information. Thereafter the format of each attribute is established for example, the transaction ID: integer, order date: Date in "YYYY-MM-DD" format etc. Upon establishing the format, the retrieving unit 210 organizes the structured data into a format that supports further operations such as feature extraction and model training. Upon performing the data definition to the retrieved data, the retrieving unit 210 is configured to pre-process the retrieved data. The pre-processing refers to the steps taken to prepare the raw data or the real-time data for use in feature extraction, model training and forecasting. The pre-processing includes, but is not limited to data cleaning, normalization, transformation. The data cleaning refers to the removing noise, duplicates, or missing values to ensure data quality. The normalization refers to the transforming of the data into a consistent format or range. The transformation includes but is not limited to, converting of data types, aggregating values or formatting of data to fit the requirements of the models. Upon pre-processing the retrieved data, the pre-processed data is stored in the storage unit 212.
[0047] Upon retrieving the data, the extracting unit 214 is configured to extract one or more features from the retrieved and pre-processed data. The one or more features are extracted from the retrieved data depending on type of training required to be performed. The one or more features refers to specific attributes, variables, or characteristics extracted from the retrieved data that are used for training forecasting models. The extraction of the one or more features depending on the type of training includes, but are not limited to, identifying relevant data columns, derived attributes and transformations. The identifying relevant data columns includes selecting the most useful fields from the raw data or real-time data. The derived attributes include creating new features based on the original data such as calculating averages or aggregations. The transformation includes applying operations such as scaling, encoding etc., to make features suitable for the model.
[0048] Upon extracting the one or more features, the configuring unit 216 is configured to configure one or more hyperparameters for the at least one of a plurality of models which are required to be trained. The one or more hyperparameters refers to the configuration parameters of the at least one of the plurality of models that are set before the training process begins. The one or more hyperparameters include, but not limited to, learning rate, number of layers or nodes, batch size, regularization parameters. The at least one of the plurality of models learns patterns/trends based on the retrieved data. In an embodiment, the model learning can happen in parallel for different models at in distributive manner and the result may come collaboratively on Graphical User Interface (GUI) or as response The one or more hyperparameters are automatically configured by the configuring unit 216, or the one or more hyperparameters are configured by the user. The automatic configuration of the one or more hyperparameters involves a specific process called hyperparameter optimization. The hyperparameter optimization is the process of automatically searching for the best set of hyperparameters that maximize the performance of the at least one of the plurality of models. In an embodiment, the request received to the system 108 through an application or microservices and receiving the result or analysis or recommendation generated by the system 108. The communication between the system 108 and the application or microservices happens via Hypertext Transfer Protocol (HTTP) (with JavaScript Object Notation/ Extensible Markup Language (JSON/XML)).
[0049] Upon configuring the one or more hyperparameters, the selecting unit 218 is configured to select a training date range and a forecasting date range for the at least one of the plurality of models. The training date range is the period of historical data selected to train the at least one of the plurality of models. For example, if a model is designed to forecast sales, the training date range could be from January 1, 2020, to December 31, 2022, i.e., the model uses sales data from this two-year period to learn and train the at least one of the plurality of models. The training date range is at least one of, dynamically set by the selecting unit 218 depending on the type of date training or, the training date range is set by the selecting unit 218 based on receiving an input from a user. The type of date training refers to the different approaches or methods used to define how the model is trained based on time-related data.
[0050] The forecasting date range is the future time period for which the trained models are used to make predictions. For example, after training on historical data, the system could predict sales for the forecasting date range of January 1, 2024, to March 31, 2024. The forecasting date range is at least one of, dynamically set by the selecting unit 218 depending on the type of forecasting the one or more events or, the forecasting date range is set by the selecting unit 218 based on receiving an input from a user. The type of forecasting includes, but is not limited to, time-series forecasting, event prediction, classification forecasting, regression forecasting. The one or more events refers to occurrences, outcomes, or phenomena that the trained models aim to predict or forecast. The one or more events includes, but are not limited to, network events, business events, system anomalies, environmental conditions.
[0051] The selecting unit 218 selects the training date range for the at least one of the plurality of models by at least one of setting a threshold of a number of datasets to be provided to the at least one of the plurality of models for training or setting, a time range to the at least one of the plurality of models for training. The setting of the threshold of the number of datasets specifies the number of data samples or instances that will be provided to the at least one of the plurality of models for training. The setting of the threshold of the number of datasets ensures that the at least one of the plurality of models is exposed to a sufficient amount of data. The setting of the time range to the at least one of the plurality of models for training refers to defining a historical period from which data will be used to train the at least one of the plurality of models. The time range is critical for models that need to learn temporal patterns, such as those used in time-series forecasting. By specifying the time range, the training process focuses on a particular period of past data, enabling the models to learn trends, seasonality, and other time-dependent relationships relevant to the forecasting task.
[0052] Further, the selecting unit 218 is configured to select the forecasting date range for the at least one of the plurality of models by setting, a threshold of a number of datasets to be forecasted by the at least one of the plurality of trained models or setting, a time range for the at least one of the plurality of models for forecasting. The setting of the threshold of the number of datasets to be forecasted by the at least one of the plurality of trained models refers to a predetermined number of data points or instances that each trained model will use to make predictions. The setting of the time range for the at least one of the plurality of models for forecasting refers to the defining of a future period over which each trained model will make predictions. The time range indicates the temporal scope of the forecast, such as predicting values or events for the next week, month, or year. By setting the time range, the models can be directed to focus on forecasting periods, allowing for targeted predictions that are relevant to the desired timeframe, whether short-term or long-term.
[0053] Upon selecting the training date range and the forecasting date range, the training unit 220 is configured to train the at least one of the plurality of models. The at least one of the plurality of models is trained with at least one of the retrieved data, the extracted one or more features, the configured one or more hyperparameters and the selected training date range. Upon training the at least one of the plurality of models, the forecasting engine 222 is configured to forecast the one or more events. The forecasting of the one or more events refers to predicting future occurrences, values, or conditions based on the models that have been trained using historical data. The one or more events are forecasted using the at least one of the plurality of trained models based on the selected forecasting date range. The forecasting engine 222, forecasts the one or more events using the at least one of the plurality of trained models, based on learnt patterns/trends of the retrieved data. The learnt patterns/trends of the retrieved data refer to the insights, correlations, and regularities that the trained models have identified from the historical data during the training process.
[0054] Upon forecasting the one or more events using each of the trained models, the recommending unit 224 is configured to recommend one or more trained models for forecasting using generated by each of the trained models. The recommending unit 224 recommends the one or more trained models for forecasting using results generated by each of the trained models, by generating, the results of the at least one of the plurality of trained models. The results include, but not limited to, accuracy values of forecasting by the at least one of the plurality of trained models, Root Mean Square Error (RMSE) values of the at least one of the plurality of trained models, training status list for the at least one of the plurality of models and forecasting status list for the at least one of the plurality of trained models. The accuracy values of forecasting by the at least one of the plurality of trained models measures the correctness of the forecasts generated by the at least one of the plurality of models, typically expressed as a percentage. The Root Mean Square Error (RMSE) values are a common metric to evaluate the prediction error of the at least one of the plurality of models. The RMSE value measures the average magnitude of forecast errors, indicating how much the predicted values deviate from the actual values. The lower RMSE values indicates the better performance. The training status list for the at least one of the plurality of models indicates the completion status of the training process for the at least one of the plurality of models, including whether training was successfully completed or if there were issues. The forecasting status list for the at least one of the plurality of trained models provides the status of the forecasting task for each trained model, detailing whether forecasting was completed successfully and any errors or issues encountered during the process.
[0055] Further, the recommending unit 224 recommends the one or more trained models for forecasting using the results generated by each of the trained models, by recommending, the one or more trained models based on the generated results. The generated results include, but are not limited to, the accuracy values of forecasting by the at least one of the plurality of trained models and the RMSE values of the at least one of the plurality of trained models. In an embodiment, the recommending unit 224 sorts the one or more trained models based on at least one of the generated indicators before recommending them.
[0056] Therefore, the system 108 recommends the appropriate forecasting logic suited for the given data source. Further, the system 108 accurately predicts the future events based on accuracy and RMSE values of each model. Further the system 108 enhances the user experience. Further the system 108 helps in optimizing the accuracy of forecasts and streamline the model selection process. Further, the system 108 reduces the manual effort required in identifying the most suitable models for a given forecasting scenario.
[0057] FIG. 3 is an exemplary block diagram of an architecture 300 of the system 108 for recommending the models for forecasting, according to one or more embodiments of the present invention.
[0058] The architecture 300 includes the one or more data sources 302, a load balancer (LB) 304, and a processing hub 306. The processing hub 306 includes data preprocessor 308, a model selection module 310, a model training unit 312, a model output display unit 314, forecasting unit 316. The forecasting unit 316 is communicably coupled with the user interface 206.
[0059] In an embodiment, the data from the one or more data sources 302 is transmitted to the LB 304. The one or more data sources 302 include, but not limited to, the HDFS, the NAS, the streaming engine, the HTTP2. Upon receiving the data from the one or more data sources 302, the LB 304 distributes or redirects bulk requests across the processing hub 306.
[0060] Uon receiving the data from the LB 304, the data preprocessor 308 pre-processes the received data. The data preprocessor 308 normalize and clean the data. The data normalization is the process of reorganizing data within the database 208 so that users can utilize it for further queries and analysis. The data cleaning is the process of fixing or removing incorrect, corrupted, incorrectly formatted, duplicate, or incomplete data within the dataset. For example, the data preprocessor 308 removes the null value, duplicate data or irrelevant data from the data received from the one or more data sources 302.
[0061] Upon preprocessing the data, the model selection module 310 selects at least one of the models from the plurality of models available. The at least one of the plurality of models are the models that would be deployed to forecast future events. In an embodiment, the future events are at least one of anomalies or issues in the network 106.
[0062] Upon selecting the at least one of the models from the plurality of models, the model training unit 312 is configured to train the selected model from the plurality of models. The selected model from the plurality of models is trained with the historical analyzed data, the extracted features from the historical data and configured hyper-parameters. The selected model from the plurality of models learns the trends, patterns and behaviors based on the received data. Upon training the forecasting model, the model training unit 312 forecasts the one or more events using the selected model from the plurality of models.
[0063] Upon training the selected model from the plurality of models and the forecasting the one or more events, the model output display unit 314 displays the output forecasted by the selected model from the plurality of models. Subsequent to displaying the output, the forecasting unit 316 is configured to provide the accuracy value and the RMSE (Root Mean Square Error) value which are used to evaluate the quality of predictions. Based on the accuracy and the RMSE value of the forecasting models, the processing hub 306 recommends the appropriate model for the user.
[0064] Thereafter, the forecasting unit 316 transmits the accuracy value and the RMSE values to the user interface 206. The user interface 206 provides the visual representation of the accuracy value of the at least one of the plurality of models to the users. Further, depending on the visual representation the user selects the appropriate model to forecast the future events.
[0065] FIG. 4 is a flow diagram for recommending the models for forecasting according to one or more embodiments of the present invention.
[0066] At step 402, retrieving the data from the one or more data sources. The one or more data sources 302 include, but not limited to, the HDFS, the NAS, the streaming engine, the HTTP2.
[0067] At step 404, upon retrieving the data from the one or more data sources 302, the data definition is performed on the retrieved data by adding static or dynamic field to the data. In particular, the fields may be added in order to make the data into a logical format. The static Fields are pre-defined and remain constant for all entries or records in the dataset. For example, retrieving sales transaction data from a database. The sales transaction data includes transaction ID, product ID, quantity sold, sales amount. Further, the static field called ‘store location’ can be added. The dynamic fields are generated based on the content or context of the retrieved data. The dynamic fields are created in real time depending on the characteristics of the data, and they might change depending on the incoming data's attributes, values, or external factors. For example, retrieving user behavior data from a web application, which includes user ID, page visited, time spent on page. The dynamic field called ‘engagement level’ is created based on the time spent on page.
[0068] At step 406, upon performing the data definition on the retrieved data, the retrieved data is preprocessed. The preprocessing of the data includes normalization and cleaning of the data to train the at least one of the plurality of model. The data cleaning is the process of identifying and correcting (or removing) inaccurate, incomplete, or irrelevant data. For example, if a transaction is logged twice, one of the duplicate entries is deleted, if a record shows a price of -50 for a product, it can either be corrected (if known) or removed from the dataset. The data normalization adjusts the range and scale of numerical data, ensuring that different features are comparable and don't disproportionately influence the model’s training.
[0069] At step 408, upon preprocessing the preprocessed data is saved in the storage unit 212 for future use.
[0070] At step 410, upon saving the preprocessed data in the storage unit 212, the one or more features are extracted from the retrieved data. The one or more features are extracted from the retrieved data depending on type of training required to be performed
[0071] At step 412, upon extracting the one or more features of the retrieved data, the one or more hyperparameters are configured for the at least one of the plurality of models which are required to be trained. The one or more hyperparameters are automatically configured by the configuring unit 216, or the one or more hyperparameters are configured by the user.
[0072] At step 414, upon configuring the one or more hyperparameters, the training date range and the forecasting date range are selected. The training date range is selected based on the type of data training or input from the user. Similarly, the forecasting date range is selected based on the type of forecasting the one or more events or input form the user.
[0073] At step 416, subsequently, the each of plurality of models are trained by using the retrieved data, the one or more extracted features, the configured one or more hyperparameters and the selected training date range. The at least one of the plurality of models learns patterns/trends based on the retrieved data.
[0074] At step 418, upon training the at least one of the plurality of models, the one or more events are forecasted using the at least one of the plurality of trained models based on the selected forecasting date range. Based on learnt patterns/trends of the retrieved data, the one or more events are forecasted using at the least one of the plurality of trained models. In an embodiment, the results generated by each of the trained models help in recommending the one or more trained models for forecasting. The results include but not limited to, the accuracy values of forecasting by the at least one of the plurality of trained models, the RMSE values of the at least one of the plurality of trained models, the training status list for the at least one of the plurality of models and the forecasting status list for the at least one of the plurality of trained models.
[0075] At step 420, subsequently, the trained models are forecasted to predict future events based on the results generated by the at least one of the plurality of trained models and provides the suitable output.
[0076] FIG. 5 is a flow diagram of a method 500 for recommending the models for forecasting according to one or more embodiments of the present invention. For the purpose of description, the method 500 is described with the embodiments as illustrated in FIG. 2 and should nowhere be construed as limiting the scope of the present disclosure.
[0077] At step 502, the method 500 includes the step of retrieving the data from one or more data sources 302 by the retrieving unit 210. The one or more data sources 302 include, but not limited to, Hadoop Distributed File System (HDFS), Non-Access Stratum (NAS), streaming engine, Hypertext Transfer Protocols (HTTP2). The retrieving unit 210 retrieves the data from the one or more data sources 302 by performing the data definition on the retrieved data. Upon performing the data definition, the retrieved unit 210 is configured to preprocess the retrieved data and store the preprocessed data in the storage unit 212.
[0078] At step 504, the method 500 includes the step of extracting the one or more features from the retrieved and pre-processed data by the extracting unit 214. The one or more features are extracted from the retrieved data depending on type of training required to be performed.
[0079] At step 506, the method 500 includes the step of configuring the one or more hyperparameters for the at least one of a plurality of models which are required to be trained by the configuring unit 216. The one or more hyperparameters are automatically configured by the configuring unit 216, or the one or more hyperparameters are configured by the user. The at least one of the plurality of models learns patterns/trends based on the retrieved data.
[0080] At step 508, the method 500 includes the step of selecting the training date range and the forecasting date range for the at least one of the plurality of models by the selecting unit 218. The selecting unit 218 selects the training date range for the at least one of the plurality of models by at least one of setting the threshold of the number of datasets to be provided to the at least one of the plurality of models for training or setting the time range to the at least one of the plurality of models for training. The training date range is at least one of, dynamically set by the selecting unit 218 depending on the type of data training or, the training date range is set by the selecting unit 218 based on receiving an input from a user. Further, the selecting unit 218 selects the forecasting date range for the at least one of the plurality of models by at least one of the setting the threshold of the number of datasets to be forecasted by the at least one of the plurality of trained models or setting the time range for the at least one of the plurality of models for forecasting. The forecasting date range is at least one of, dynamically set by the selecting unit 218 depending on the type of forecasting the one or more events or, the forecasting date range is set by the selecting unit 218 based on receiving an input from a user.
[0081] At step 510, the method 500 includes the step of training the at least one of the plurality of models with at least one of, the retrieved data, the extracted one or more features, the configured one or more hyperparameters and the selected training date range by the training unit 220.
[0082] At step 512, the method 500 includes the step of forecasting the one or more events using the at least one of the plurality of trained models based on the selected forecasting date range by the forecasting engine 222. The forecasting engine 222 forecasts the one or more events using the at least one of the plurality of trained models, based on learnt patterns/trends of the retrieved data.
[0083] At step 514, the method 500 includes the step of recommending the one or more trained models for forecasting using results generated by each of the trained models by the recommending unit 224. The recommending unit 224 recommends the one or more trained models for forecasting using results generated by each of the trained models, by generating the results of the at least one of the plurality of trained models. The results include, but not limited to, the accuracy values of forecasting by at least one of the plurality of trained models, the RMSE values of the at least one of the plurality of trained models, the training status list for the at least one of the plurality of models and the forecasting status list for the at least one of the plurality of trained models. Further, the recommending unit 224 recommends the one or more trained models for forecasting using results generated by each of the trained models, by recommending, the one or more trained models based on the generated results including at least one of, the accuracy values of forecasting by the at least one of the plurality of trained models and the RMSE values of the at least one of the plurality of trained models.
[0084] The present invention further discloses a non-transitory computer-readable medium having stored thereon computer-readable instructions. The computer-readable instructions are executed by the processor 202. The processor 202 is configured to retrieve the data from the one or more data sources 302. The processor 202 is further configured to extract the one or more features from the retrieved and pre-processed data. The processor 202 is further configured to configure the one or more hyperparameters for the at least one of the plurality of models which are required to be trained. The processor 202 is further configured to select the training date range and the forecasting date range for the at least one of the plurality of models. The processor 202 is further configured to train, the at least one of the plurality of models with at least one of, the retrieved data, the extracted one or more features, the configured one or more hyperparameters and the selected training date range. The processor 202 is further configured to forecast the one or more events using the at least one of the plurality of trained models based on the selected forecasting date range. The processor 202 is further configured to recommend one or more trained models for forecasting using results generated by each of the trained models.
[0085] A person of ordinary skill in the art will readily ascertain that the illustrated embodiments and steps in description and drawings (FIG.1-5) 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 and spirit of the disclosed embodiments.
[0086] The present disclosure incorporates technical advancement of recommending the appropriate forecasting logic suited for the given data source. Further, the present invention accurately predicts the future events based on accuracy and RMSE values of each model. Further the present invention enhances the user experience. Further the present invention helps in optimizing the accuracy of forecasts and streamline the model selection process. Further, the present invention reduces the manual effort required in identifying the most suitable models for a given forecasting scenario.
[0087] The present invention offers multiple advantages over the prior art and the above listed are a few examples to emphasize on some of the advantageous features. The listed advantages are to be read in a non-limiting manner.

REFERENCE NUMERALS

[0088] Environment- 100
[0089] User Equipment (UE)- 102
[0090] Server- 104
[0091] Network- 106
[0092] System -108
[0093] Processor- 202
[0094] Memory- 204
[0095] User Interface- 206
[0096] Database- 208
[0097] Retrieving Unit - 210
[0098] Storage Unit- 212
[0099] Extracting unit- 214
[00100] Configuring Unit- 216
[00101] Selecting Unit- 218
[00102] Training Unit- 220
[00103] Forecasting hub- 222
[00104] Recommending Unit- 224
[00105] One or more data sources- 302
[00106] Load balancer (LB)-304
[00107] Processing hub- 306
[00108] Data preprocessor - 308
[00109] Model selection module- 310
[00110] Model Training unit- 312
[00111] Model output display unit- 314
[00112] Forecasting unit- 316
,CLAIMS:CLAIMS:
We Claim:
1. A method (500) for recommending models for forecasting, the method (500) comprising the steps of:
retrieving, by one or more processors (202), data from one or more data sources (302);
extracting, by the one or more processors (202), one or more features from the retrieved and pre-processed data;
configuring, by the one or more processors (202), one or more hyperparameters for at least one of a plurality of models which are required to be trained;
selecting, by the one or more processors (202), a training date range and forecasting date range for the at least one of the plurality of models;
training, by the one or more processors (202), the at least one of the plurality of models with at least one of, the retrieved data, the extracted one or more features, the configured one or more hyperparameters and the selected training date range;
forecasting, by the one or more processors (202), one or more events using at least one of the plurality of trained models based on the selected forecasting date range; and
recommending, by the one or more processors (202), one or more trained models for forecasting using results generated by at least one of the trained models.

2. The method (500) as claimed in claim 1, wherein the one or more data sources (302) include, but not limited to, Hadoop Distributed File System (HDFS), Non-Access Stratum (NAS), streaming engine, Hypertext Transfer Protocols (HTTP2).

3. The method (500) as claimed in claim 1, wherein the step of, retrieving, data from one or more data sources (302), further includes the steps of:
performing, by the one or more processors (202), data definition on the retrieved data;
pre-processing, by the one or more processors (202), the retrieved data; and
storing, by the one or more processors (202), the pre-processed data in a storage unit (212).

4. The method (500) as claimed in claim 1, wherein the one or more features are extracted from the retrieved data depending on type of training required to be performed.

5. The method (500) as claimed in claim 1, wherein the step of, selecting, a training date range for at least one of the plurality of models includes at least one of the steps of:
setting, by the one or more processors (202), a threshold of a number of datasets to be provided to the at least one of the plurality of models for training; or
setting, by the one or more processors (202), a time range to the at least one of the plurality of models for training.

6. The method (500) as claimed in claim 1, wherein the training date range is at least one of, dynamically set by the one or more processors (202) depending on the type of data training or, the training date range is set by the one or more processors (202) based on receiving an input from a user.

7. The method (500) as claimed in claim 1, wherein the step of, selecting, a forecasting date range for the at least one of the plurality of models includes at least one of the steps of:
setting, by the one or more processors (202), a threshold of a number of datasets to be forecasted by the at least one of the plurality of trained models; or
setting, by the one or more processors (202), a time range for the at least one of the plurality of models for forecasting.

8. The method (500) as claimed in claim 1, wherein the forecasting date range is at least one of, dynamically set by the one or more processors (202) depending on the type of forecasting the one or more events or, the forecasting date range is set by the one or more processors (202) based on receiving an input from a user.

9. The method (500) as claimed in claim 1, wherein the one or more hyperparameters are automatically configured by the one or more processors (202), or the one or more hyperparameters are configured by the user.

10. The method (500) as claimed in claim 1, wherein the one or more processors (202) enables the at least one of the plurality of models to learn patterns/trends based on the retrieved data.

11. The method (500) as claimed in claim 1, wherein the one or more processors (202), forecasts the one or more events using the at least one of the plurality of trained models, based on learnt patterns/trends of the retrieved data.

12. The method (500) as claimed in claim 1, wherein the step of, recommending, one or more trained models for forecasting using results generated by the at least one of the trained models, includes the steps of:
generating, by the one or more processors (202), the results of the at least one of the plurality of trained models, wherein the results include ,but not limited to, accuracy values of forecasting by the at least one of the plurality of trained models, Root Mean Square Error (RMSE) values of the at least one of the plurality of trained models, training status list for the at least one of the plurality of models and forecasting status list for the at least one of the plurality of trained models; and
recommending, by the one or more processors (202), the one or more trained models based on the generated results including at least one of, the accuracy values of forecasting by the at least one of the plurality of trained models and the RMSE values of the at least one of the plurality of trained models.

13. A system (108) for recommending models for forecasting, the system (108) comprising:
a retrieving unit (210), configured to, retrieve, data from one or more data sources (302);
an extracting unit (214), configured to, extract, one or more features from the retrieved and pre-processed data;
a configuring unit (216), to configure, one or more hyperparameters for at least one of a plurality of models which are required to be trained;
a selecting unit (218), configured to, select, a training date range and a forecasting date range for at least one of the plurality of models;
a training unit (220), configured to, train, the at least one of the plurality of models with at least one of, the retrieved data, the extracted one or more features, the configured one or more hyperparameters and the selected training date range;
a forecasting engine (222), configured to, forecast, one or more events using at least one of the plurality of trained models based on the selected forecasting date range; and
a recommending unit (224), configured to, recommend, one or more trained models for forecasting using results generated by the at least one of the trained models.

14. The system (108) as claimed in claim 13, wherein the one or more data sources (302) include, but not limited to, Hadoop Distributed File System (HDFS), Non-Access Stratum (NAS), streaming engine, Hypertext Transfer Protocols (HTTP2).

15. The system (108) as claimed in claim 13, wherein the retrieving unit (210), is further configured to:
perform, data definition on the retrieved data;
pre-process, the retrieved data; and
store, the pre-processed data in a storage unit (212).

16. The system (108) as claimed in claim 13, wherein the one or more features are extracted from the retrieved data depending on type of training required to be performed.

17. The system (108) as claimed in claim 13, wherein the selecting unit (218), selects, the training date range for the at least one of the plurality of models by at least one of:
setting, a threshold of a number of datasets to be provided to the at least one of the plurality of models for training; or
setting, a time range to the at least one of the plurality of models for training.

18. The system (108) as claimed in claim 13, wherein the training date range is at least one of, dynamically set by the selecting unit (218) depending on the type of data training or, the training date range is set by the selecting unit based on receiving an input from a user.

19. The system (108) as claimed in claim 13, wherein the selecting unit (218) selects, the forecasting date range for the at least one of the plurality of models by:
setting, a threshold of a number of datasets to be forecasted by the at least one of the plurality of trained models; or
setting, a time range for the at least one of the plurality of models for forecasting.

20. The system (108) as claimed in claim 13, wherein the forecasting date range is at least one of, dynamically set by the selecting unit (218) depending on the type of forecasting the one or more events or, the forecasting date range is set by the selecting unit based on receiving an input from a user.

21. The system (108) as claimed in claim 13, wherein the one or more hyperparameters are automatically configured by the configuring unit (216), or the one or more hyperparameters are configured by the user.

22. The system (108) as claimed in claim 13, wherein the at least one of the plurality of models learns patterns/trends based on the retrieved data.

23. The system (108) as claimed in claim 13, wherein the forecasting engine (222), forecasts the one or more events using the at least one of the plurality of trained models, based on learnt patterns/trends of the retrieved data.

24. The system (108) as claimed in claim 13, wherein the recommending unit (224), recommends, one or more trained models for forecasting using results generated by each of the trained models, by:
generating, the results of the at least one of the plurality of trained models, wherein the results include, but not limited to, accuracy values of forecasting by the at least one of the plurality of trained models, Root Mean Square Error (RMSE) values of the at least one of the plurality of trained models, training status list for the at least one of the plurality of models and forecasting status list for the at least one of the plurality of trained models; and
recommending, the one or more trained models based on the generated results including at least one of, the accuracy values of forecasting by the at least one of the plurality of trained models and the RMSE values of the at least one of the plurality of trained models.