DESC:RESERVATION OF RIGHTS
[001] A portion of the disclosure of this patent document contains material, which is subject to intellectual property rights such as, but are not limited to, copyright, design, trademark, Integrated Circuit (IC) layout design, and/or trade dress protection, belonging to Jio Platforms Limited (JPL) or its affiliates (hereinafter referred as owner). The owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in the Patent and Trademark Office patent files or records, but otherwise reserves all rights whatsoever. All rights to such intellectual property are fully reserved by the owner.

FIELD OF DISCLOSURE
[002] The embodiments of the present disclosure generally relate to generating recommendations for e-commerce business. In particular, the present disclosure relates to a system and a method for recommending products to users of an e-commerce platform that includes businesses and merchants/distributors.

BACKGROUND OF DISCLOSURE
[003] The following description of related art is intended to provide background information pertaining to the field of the disclosure. This section may include certain aspects of the art that may be related to various features of the present disclosure. However, it should be appreciated that this section be used only to enhance the understanding of the reader with respect to the present disclosure, and not as admissions of prior art.
[004] In general, product recommendations help Business-to-Business (B2B) merchants deliver more relevant product recommendations to individual buyers, increase conversion rates and average order values, and automate once tedious and time-consuming merchandising tasks. By freeing up resources, merchandisers can work on more critical areas like analytics, user experiences, search management, and more.
[005] Recommendation engines are a backbone of most modern commerce applications and play a huge role in customer engagement as well as in upselling and cross selling of the products. However, since most widely known use cases of e-commerce platforms are Business-to-Customer (B2C), most of available research also caters to B2C domain and Business-to-Business (B2B) domain is not researched/reported on sufficiently. Typically, B2B end users behave differently from B2C end users, since for businesses, a lot of purchase is related to re-ordering of known old products with some small degree of exploration. Currently available mechanisms for product recommendation relate to recommending products based on a distributed personalized recommendation methods and are inefficient in terms of time, cost, and performance.
[006] There is, therefore, a need in the art to provide a system and method for recommending products to users of an e-commerce platform that includes businesses and merchants/distributors to overcome the shortcomings of the existing systems and methods.

OBJECTS OF THE PRESENT DISCLOSURE
[007] Some of the objects of the present disclosure, which at least one embodiment herein satisfies are as listed herein below.
[008] It is an object of the present disclosure to provide personalized product recommendations to merchants and/or distributors.
[009] It is an object of the present disclosure to provide product recommendations to the merchants and/or distributors in a Business-to-Business (B2B) scenario.
[0010] It is an object of the present disclosure to provide product recommendations while improving quality of the recommendations made by predicting both new products that may be bought and also the products that were previously bought by users, thereby leading to significant uplifting of the merchant’s buying experience.
[0011] It is an object of the present disclosure to transform shopping experience of the merchants.
[0012] It is an object of the present disclosure to provide the recommendations to ease or reduce manual searching of the products and minimize user effort.
[0013] It is an object of the present disclosure to determine price/margin sensitivity that enables to tackle out of stock situations as well as provides additional information to improve recommendations of the products.

SUMMARY
[0014] This section is provided to introduce certain objects and aspects of the present disclosure in a simplified form that are further described below in the detailed description. This summary is not intended to identify the key features or the scope of the claimed subject matter.
[0015] In an aspect, the present disclosure relates to a system for providing personalised recommendations. The system includes one or more processors, and a memory operatively coupled to the one or more processors, where the memory includes processor-executable instructions, which on execution, cause the one or more processors to predict user preference of a product by collecting historical preferences of a group of users, predict, via a trained model, reordering of the product and ordering of a new product based on the user preference of the product, determine an exploration rate between the new product and a reordered product based on a plurality of regression factors, and provide one or more personalized recommendations corresponding to the new product and the reordered product to an end user based on the exploration rate.
[0016] In an embodiment, the memory includes processor-executable instructions, which on execution, may cause the one or more processors to determine price and margin sensitivity of a merchant or distributor with respect to a category of the product.
[0017] In an embodiment, the memory includes processor-executable instructions, which on execution, may cause the one or more processors to determine a change in price and margin of the reordered product and replace the reordered product with an alternative product based on the determined change.
[0018] In an embodiment, the processor may be to predict the reordering of the product and the ordering of the new product by being configured to train a respective model for each combination of a unique area and a unique product, determine, via the trained model, whether the product is ordered multiple times or the product is ordered for a first time based on the historical preferences of the group of users, predict the reordering of the product in response to determining that the product is ordered multiple times, and predict the ordering of the new product in response to determining that the product is ordered for the first time.
[0019] In an embodiment, the memory includes processor-executable instructions, which on execution, may cause the one or more processors to optimize a balance between the reordering of the product and the ordering of the new product based on the exploration rate.
[0020] In an embodiment, the memory includes processor-executable instructions, which on execution, may cause the one or more processors to identify one or more similar products by creating a hybrid model, wherein the hybrid model may be created by combining a Term Frequency-Inverse Document Frequency (TF-IDF) vector-based similarity and a Bidirectional Encoder Representation from Transformers (BERT) vector-based similarity, and provide the one or more personalized recommendations corresponding to the one or more similar products to the end user using the hybrid model.
[0021] In an embodiment, the memory includes processor-executable instructions, which on execution, may cause the one or more processors to provide the one or more personalized recommendations corresponding to an alternative product based on the product ordered by the end user being out of stock.
[0022] In an embodiment, the plurality of regression factors may include at least one of an age on a distribution platform, number of days after last purchase of the at least one product, an average amount spent per transaction, an amount spent in last transaction, and an average number of unique products bought per transaction.
[0023] In another aspect, the present disclosure relates to a method for providing personalised recommendations. The method includes predicting, by a processor associated with a system, user preference of a product by collecting historical preferences of a group of users, predicting, by the processor, via a trained model, reordering of the product and ordering of a new product based on the user preference of the product, determining, by the processor, an exploration rate between the new product and a reordered product based on a plurality of regression factors, and providing, by the processor, one or more personalized recommendations corresponding to the new product and the reordered product to an end user based on the exploration rate.
[0024] In an embodiment, the method may include determining, by the processor, price and margin sensitivity of a distributor with respect to a category of the product.
[0025] In an embodiment, the method may include determining, by the processor, a change in price and margin of the reordered product and replacing, by the processor, the reordered product with an alternative product based on the determined change.
[0026] In an embodiment, predicting, by the processor via the trained model, the reordering of the product and the ordering of the new product may include training, by the processor, a respective model for each combination of a unique area and a unique product, determining, by the processor, via the trained model, whether the product is ordered multiple times or the product is ordered for the first time based on the historical preferences of the group of users, predicting, by the processor, the reordering of the product in response to determining that the product is ordered multiple times, and predicting, by the processor, the ordering of the new product in response to determining that the product is ordered for the first time.
[0027] In an embodiment, the method may include optimizing, by the processor, a balance between the reordering of the product and the ordering of the new product based on the exploration rate.
[0028] In an embodiment, the method may include identifying, by the processor, one or more similar products by creating a hybrid model, wherein the hybrid model may be created by combining a Term Frequency-Inverse Document Frequency (TF-IDF) vector-based similarity and a Bidirectional Encoder Representation from Transformers (BERT) vector-based similarity, and providing, by the processor, the one or more personalized recommendations corresponding to the one or more similar products to the end user using the hybrid model.
[0029] In an embodiment, the method may include determining, by the processor, that the product ordered by the end user is out of stock, and providing, by the processor, the one or more personalized recommendations corresponding to an alternative product in response to determining that the product ordered by the end user is out of stock.
[0030] In an embodiment, the plurality of regression factors may include at least one of an age on a distribution platform, number of days after last purchase of the at least one product, an average amount spent per transaction, an amount spent in last transaction, and an average number of unique products bought per transaction.
[0031] In another aspect, the present disclosure relates to a user equipment. The user equipment includes one or more processors, and a memory operatively coupled to the one or more processors, wherein the memory includes processor-executable instructions, which on execution, cause the one or more processors to receive one or more personalized recommendations from a system via a wireless network. The one or more processors are communicatively coupled with the system, and wherein the system is configured to predict user preference of a product by collecting historical preferences of a group of users, predict, via a trained model, reordering of the product and ordering of a new product based on the user preference of the product, determine an exploration rate between the new product and a reordered product based on a plurality of regression factors, and provide one or more personalized recommendations corresponding to the new product and the reordered product to an end user associated with the user equipment based on the exploration rate.

BRIEF DESCRIPTION OF THE DRAWINGS
[0032] The accompanying drawings, which are incorporated herein, and constitute a part of this invention, 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 invention. 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 invention of such drawings includes the invention of electrical components, electronic components or circuitry commonly used to implement such components.
[0033] The diagrams are for illustration only, which thus is not a limitation of the present disclosure, and wherein:
[0034] FIG. 1 illustrates an exemplary network architecture 100 in which or with which embodiments of the present disclosure may be implemented.
[0035] FIG. 2 illustrates an exemplary block diagram 200 of a Business-to-Business (B2B) recommendation system, in accordance with an embodiment of the present disclosure.
[0036] FIG. 3 illustrates an exemplary architecture 300 for implementing the B2B recommendation system, in accordance with an embodiment of the present disclosure.
[0037] FIG. 4 illustrates an exemplary flow diagram 400 for generating recommendations using a matrix factorization engine, in accordance with an embodiment of the present disclosure.
[0038] FIG. 5 illustrates an exemplary flow diagram 500 for implementing a two-level reorder prediction engine, in accordance with an embodiment of the present disclosure.
[0039] FIG. 6 illustrates an exemplary flow diagram 600 for implementing a price/margin sensitivity computation engine, in accordance with an embodiment of the present disclosure.
[0040] FIG. 7 illustrates an exemplary flow diagram 700 for implementing an exploration rate modelling engine, in accordance with an embodiment of the present disclosure.
[0041] FIG. 8 illustrates an exemplary flow diagram 800 for implementing a product similarity engine, in accordance with an embodiment of the present disclosure.
[0042] FIG. 9 illustrates an exemplary flow diagram 900 for implementing a distribution engine, in accordance with an embodiment of the present disclosure.
[0043] FIG. 10 illustrates an exemplary flow diagram of a method 1000 for providing personalized recommendations to an end user, in accordance with an embodiment of the present disclosure.
[0044] FIG. 11 illustrates an exemplary computer system 1100 in which or with which embodiments of the present disclosure may be implemented.
[0045] The foregoing shall be more apparent from the following more detailed description of the disclosure.

DETAILED DESCRIPTION
[0046] In the following description, for the purposes of explanation, various specific details are set forth in order to provide a thorough understanding of embodiments of the present disclosure. It will be apparent, however, that embodiments of the present disclosure may be practiced without these specific details. Several features described hereafter can each be used independently of one another or with any combination of other features. An individual feature may not address all of the problems discussed above or might address only some of the problems discussed above. Some of the problems discussed above might not be fully addressed by any of the features described herein.
[0047] The ensuing description provides exemplary embodiments only, and is not intended to limit the scope, applicability, or configuration of the disclosure. Rather, the ensuing description of the exemplary embodiments will provide those skilled in the art with an enabling description for implementing an exemplary embodiment. It should be understood that various changes may be made in the function and arrangement of elements without departing from the spirit and scope of the disclosure as set forth.
[0048] Specific details are given in the following description to provide a thorough understanding of the embodiments. However, it will be understood by one of ordinary skill in the art that the embodiments may be practiced without these specific details. For example, circuits, systems, networks, processes, and other components may be shown as components in block diagram form in order not to obscure the embodiments in unnecessary detail. In other instances, well-known circuits, processes, algorithms, structures, and techniques may be shown without unnecessary detail in order to avoid obscuring the embodiments.
[0049] Also, it is noted that individual embodiments may be described as a process which is depicted as a flowchart, a flow diagram, a data flow diagram, a structure diagram, or a block diagram. Although a flowchart may describe the operations as a sequential process, many of the operations can be performed in parallel or concurrently. In addition, the order of the operations may be re-arranged. A process is terminated when its operations are completed but could have additional steps not included in a figure. A process may correspond to a method, a function, a procedure, a subroutine, a subprogram, etc. When a process corresponds to a function, its termination can correspond to a return of the function to the calling function or the main function.
[0050] The word “exemplary” and/or “demonstrative” is used herein to mean serving as an example, instance, or illustration. For the avoidance of doubt, the subject matter disclosed herein is not limited by such examples. In addition, any aspect or design described herein as “exemplary” and/or “demonstrative” is not necessarily to be construed as preferred or advantageous over other aspects or designs, nor is it meant to preclude equivalent exemplary structures and techniques known to those of ordinary skill in the art. Furthermore, to the extent that the terms “includes,” “has,” “contains,” and other similar words are used in either the detailed description or the claims, such terms are intended to be inclusive in a manner similar to the term “comprising” as an open transition word without precluding any additional or other elements.
[0051] Reference throughout this specification to “one embodiment” or “an embodiment” or “an instance” or “one instance” means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present disclosure. Thus, the appearances of the phrases “in one embodiment” or “in an embodiment” in various places throughout this specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
[0052] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. As used herein, the singular forms “a”, “an” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms “comprises” and/or “comprising,” when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof. As used herein, the term “and/or” includes any and all combinations of one or more of the associated listed items.
[0053] In prevalent times, recommendation engines are a backbone of most e-commerce applications and play a huge role in customer engagement as well as in product upselling and cross selling. However, Business-to-Business (B2B) end users behave differently from Business-to-Consumer (B2C) end users, since for businesses, a lot of purchase is done in terms of reordering of known old products with lesser degree of exploration. The disclosure facilitates to provide an optimized balance for predicting new and old products for each merchant.
[0054] The present disclosure optimizes the balance between predicting new and old products for each merchant by predicting ordering of the new products and re-ordering of the old products both separately and combining them in a proportion defined as an exploration rate. The present disclosure learns the exploration rate for each merchant from buying history of each merchant. Further, the present disclosure predicts reordering of the products by the merchants by combining elements of Hierarchical Time Series Modelling with learned heuristics. Furthermore, the present disclosure proposes an experimental framework to calculate price sensitivity and margin sensitivity to improve the recommendations and have alternatives when the platform runs out of stock for any product.
[0055] Therefore, embodiments of the present disclosure relate to generating and providing relevant recommendations that combines new buying and rebuying of the products, thereby, improving customer experience and engagement during purchase of the product. In an embodiment, the present disclosure provides alternative recommendations when the platform runs out of stock for any product.
[0056] In an embodiment, the present disclosure performs reorder prediction to predict what products are likely to be ordered again by the merchants. In an embodiment, the present disclosure determines price/margin sensitivity of the merchant with respect to a category of the product. In an embodiment, the present disclosure measures the exploration rate of the merchant, i.e., distribution between the new product and the previously ordered products at each transaction.
[0057] Certain terms and phrases have been used throughout the disclosure and will have the following meanings in the context of the ongoing disclosure.
[0058] The term “B2B” may refer to Business-to-Business which is a form of transaction between businesses, such as one involving a manufacturer and wholesaler, or a wholesaler and a retailer.
[0059] The term “B2C” may refer to Business-to-customer which is a process of selling products and services directly between the business and customers who are the end users of the products or services.
[0060] The term “recommendation” may refer to suggestions provided to the end users regarding new products, previously ordered products, and alternative products.
[0061] The term “exploration rate” may refer to a distribution between a new product and previously ordered products at each transaction.
[0062] Various embodiments of the present disclosure will be explained in detail with reference to FIGs. 1-11.
[0063] FIG. 1 illustrates an exemplary network architecture (100) in which or with which embodiments of the present disclosure may be implemented.
[0064] Referring to FIG. 1, the network architecture (100) may include one or more user equipments (104-1, 104-2…104-N) associated with one or more users (102-1, 102-2…102-N) in an environment. A person of ordinary skill in the art will understand that one or more users (102-1, 102-2…102-N) may be individually referred to as the user (102) and collectively referred to as the users (102). Similarly, a person of ordinary skill in the art will understand that one or more user equipments (104-1, 104-2…104-N) may be individually referred to as the user equipment (104) and collectively referred to as the user equipment (104). A person of ordinary skill in the art will appreciate that the terms “computing device(s)” and “user equipment” may be used interchangeably throughout the disclosure. Although three user equipments (104) are depicted in FIG. 1, however any number of the user equipments (104) may be included without departing from the scope of the ongoing description.
[0065] In an embodiment, the user equipment (104) may include smart devices operating in a smart environment, for example, an Internet of Things (IoT) system. In such an embodiment, the user equipment (104) may include, but is not limited to, smart phones, smart watches, smart sensors (e.g., mechanical, thermal, electrical, magnetic, etc.), networked appliances, networked peripheral devices, networked lighting system, communication devices, networked vehicle accessories, networked vehicular devices, smart accessories, tablets, smart television (TV), computers, smart security system, smart home system, other devices for monitoring or interacting with or for the users (102) and/or entities, or any combination thereof.
[0066] A person of ordinary skill in the art will appreciate that the user equipment (104) may include, but is not limited to, intelligent, multi-sensing, network-connected devices, that can integrate seamlessly with each other and/or with a central server or a cloud-computing system or any other device that is network-connected.
[0067] In an embodiment, the user equipment (104) may include, but is not limited to, a handheld wireless communication device (e.g., a mobile phone, a smart phone, a phablet device, and so on), a wearable computer device(e.g., a head-mounted display computer device, a head-mounted camera device, a wristwatch computer device, and so on), a Global Positioning System (GPS) device, a laptop computer, a tablet computer, or another type of portable computer, a media playing device, a portable gaming system, and/or any other type of computer device with wireless communication capabilities, and the like. In an embodiment, the user equipment (104) may include, but is not limited to, any electrical, electronic, electro-mechanical, or an equipment, or a combination of one or more of the above devices such as 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, wherein the user equipment (104) may include one or more in-built or externally coupled accessories including, but not limited to, a visual aid device such as a camera, an audio aid, a microphone, a keyboard, and input devices for receiving input from the user (102) or the entity such as touch pad, touch enabled screen, electronic pen, and the like.
[0068] A person of ordinary skill in the art will appreciate that the user equipment (104) may not be restricted to the mentioned devices and various other devices may be used.
[0069] Referring to FIG. 1, the user equipment (104) may communicate with a system (110), for example, a B2B recommendation system, through a network (106). In an embodiment, the network (106) may include at least one of a Fifth Generation (5G) network, or the like. The network (106) may enable the user equipment (104) to communicate with other devices in the network architecture (100) and/or with the system (110). The network (106) may include a wireless card or some other transceiver connection to facilitate this communication. In another embodiment, the network (106) may be implemented as, or include any of a variety of different communication technologies such as a wide area network (WAN), a local area network (LAN), a wireless network, a mobile network, a Virtual Private Network (VPN), the Internet, the Public Switched Telephone Network (PSTN), or the like.
[0070] In accordance with embodiments of the present disclosure, the system (110) may be designed and configured for providing personalized recommendations corresponding to products in an e-commerce platform to the end users (102). As such, the system (110) may have the capability to predict reordering of previously bought products and ordering of new products separately. Alternatively, or additionally, the system (110) may provide recommendations corresponding to alternative products when the ordered product is out of stock.
[0071] In accordance with embodiments of the present disclosure, the B2B recommendation system (110) may predict user preference of a product, such as for example, but not limited to manufacturing materials, clothing, car parts, and semiconductors by collecting historical preferences of a group of users (102). Further, the B2B recommendation system (110) may predict the reordering of the products, and the ordering of the new product via a trained model. Also, the B2B recommendation system (110) may determine an exploration rate between the new product and the reordered product based on a plurality of regression factors. The regression factors include, but not limited to, an age on a distribution platform, number of days after last purchase of the product, an average amount spent per transaction, an amount spent in last transaction, and an average number of unique products bought per transaction. Furthermore, according to various embodiments of the present disclosure, one or more personalized recommendations corresponding to the new product and the at least one reordered product may be provided to the end users (102) based on the exploration rate. In some embodiments, the one or more personalized recommendations may correspond to similar products. In some embodiments, the one or more personalized recommendations may correspond to alternative products when the ordered product is out of stock. Therefore, the present disclosure facilitates to improve the user experience and engagement while purchasing products.
[0072] Although FIG. 1 shows exemplary components of the network architecture (100), in other embodiments, the network architecture (100) may include fewer components, different components, differently arranged components, or additional functional components than depicted in FIG. 1. Additionally, or alternatively, one or more components of the network architecture (100) may perform functions described as being performed by one or more other components of the network architecture (100).
[0073] FIG. 2 illustrates an exemplary block diagram of the B2B recommendation system (200), in accordance with an embodiment of the present disclosure. It may be appreciated that the system (200) may be similar to the system (110) of FIG. 1 in its functionality.
[0074] In an embodiment, and as shown in FIG. 2, the system (200) may include one or more processors (202). The one or more processors (202) may be implemented as one or more microprocessors, microcomputers, microcontrollers, digital signal processors, central processing units, logic circuitries, and/or any devices that manipulate data based on operational instructions. Among other capabilities, the one or more processor(s) (202) may be configured to fetch and execute computer-readable instructions stored in a memory (204) of the system (200). The memory (204) may store one or more computer-readable instructions or routines, which may be fetched and executed to create or share the data units over a network service. The memory (204) may comprise any non-transitory storage device including, for example, volatile memory such as Random-Access Memory (RAM), or non-volatile memory such as Erasable Programmable Read-Only Memory (EPROM), flash memory, and the like.
[0075] In an embodiment, the system (200) may also comprise an interface(s) (206). The interface(s) (206) may comprise a variety of interfaces, for example, a variety of interfaces, for example, interfaces for data input and output devices, referred to as I/O devices, storage devices, and the like. The interface(s) (206) may facilitate communication of the system (200) with various devices coupled to it. The interface(s) (206) may also provide a communication pathway for one or more components of the system (200). Examples of such components include, but are not limited to, processing engine(s) (208) and a database (210).
[0076] In an embodiment, the processing engine(s) (208) may be implemented as a combination of hardware and programming (for example, programmable instructions) to implement one or more functionalities of the processing engine(s) (208). In examples described herein, such combinations of hardware and programming may be implemented in several different ways. For example, the programming for the processing engine(s) (208) may be processor executable instructions stored on a non-transitory machine-readable storage medium and the hardware for the one or more processors (202) may comprise a processing resource (for example, one or more processors), to execute such instructions. In the present examples, the machine-readable storage medium may store instructions that, when executed by the processing resource, implement the processing engine(s) (208). In such examples, the system (200) may comprise the machine-readable storage medium storing the instructions and the processing resource to execute the instructions, or the machine-readable storage medium may be separate but accessible to the system (200) and the processing resource. In other examples, the processing engine(s) (208) may be implemented by electronic circuitry.
[0077] In an embodiment, the database (210) may comprise data that may be either stored or generated as a result of functionalities implemented by any of the components of the processor(s) (202) or the processing engine(s) (208) or the system (200). In an exemplary embodiment, the processing engine(s) (208) may include a matrix factorization engine (212), an exploration rate prediction engine (214), a reorder prediction engine (216), a price and margin sensitivity experimentation engine (218), an item similarity engine (220), and other engines (222). The other engines (222) may further include, without limitation, a data receiving engine, a storage engine, a computing engine, or a signal generation engine. Other engine(s) (222) may supplement the functionalities of the processing engine(s) (208) or the system (200). The system (200) may be implemented using any or a combination of hardware components and software components.
[0078] In an embodiment, the matrix factorization engine (212) may use a Bayesian personalized ranking (BPR) or any similar matrix factorization algorithm to compute user and item vectors from buying history of users (e.g., 102). The BPR algorithm may define the loss function for matrix factorization. The BPR algorithm may predict user preference of the product or the item by collecting historical preferences of a group of users (102).
[0079] In an embodiment, the exploration rate prediction engine (214) may capture that different merchants’ exhibit different behaviour with regards to buying new products v/s rebuying previously bought products for each purchase. The behaviour of the merchants may be influenced by a number of factors and training of a Linear Regression Model based on the regression factors. The regression factors may include, but not be limited to, an age on a platform, a number of days since last purchase of the product, an average amount spent per transaction, an amount spent in last transaction, an average number of unique items bought per transaction, a number of unique items bought in last transaction, top products bought in last N transactions, top products bought in last transaction, and average distribution of new v/s old products in previous orders. An exploration rate may be determined by the exploration rate prediction engine (214) between the new product and the old product based on regression on the mentioned independent variables.
[0080] In an embodiment, forecasting models may be created at a merchant-product level leading to producing a number of models, i.e., Num Merchants*Num Products models. However, it may be hard to train and maintain the produced models. In addition, due to sparsity of data at the merchant-product level, the produced models may be error prone. To tackle this, the reorder prediction engine (216) may follow a hierarchical time series approach. The hierarchical time series approach facilitates to perform reorder prediction for the merchants. The hierarchical time series approach aggregates product sales at a city level and enables to remove noise as well as sparsity of the data, and therefore becomes much easier to model. By way of an example, the level may be modified to any hierarchy, for example, state, city, zip-code etc.
[0081] In an embodiment, a unique model may be provided at a city product level, i.e., a different model may be provided for each combination of a unique city and a unique product. Upon receiving the forecast for a given city-product for a specific time period (for example, a month), the forecasts may be distributed across the merchants in the city who buy the product using a distribution method based on historical buying. In the distribution method, a historical proportion is defined as a quantity of a product bought by the merchant as a proportion of total quantity bought in the city. The proportion is averaged over many time periods to remove noise. The historical proportion is determined as mentioned below.
…. eq (1)
.… eq (2)

.… eq (3)

[0082] With respect to Equation 1 (eq 1): the proportion of the merchant M for Product N(Pm,n), is defined as Quantity of N bought by merchant M divided by Sum of Quantity of N bought by all the merchants.
[0083] With respect to Equation 2 (eq 2): an averaging of P over t-k-1 time periods is determined to smoothen it out. Here k may be a constant to determine the smoothing time period.
[0084] With respect to Equation 3 (eq 3): a multiply product forecast (F) is determined by merchant proportion to get expected quantity.
[0085] In an embodiment, ranking of the recommendations is considered important since screen size and user attention is a constraint. Hence, there is a need to serve the most effective recommendations at the beginning. To rank the products, Expected Quantity: E(Q) and product margin is combined.
.…. eq (4)
Where M(m) is margin of Product M defined as MRP of M – Buying Price of M.
[0086] In an embodiment, sorting the product on ranking metric in a descending order provides an optimal rank of the recommendations.
[0087] Conventionally defined product similarity model is used at a single product level and is used by companies to discover the best price for their product. However, use case of a middleman (i.e., a platform) is substantially different since the middleman does not own control prices.
[0088] To overcome this, in an embodiment, the present disclosure provides the price and margin sensitivity experimentation engine (218) to determine an elasticity of the merchant and determine replaceability of the product, thereby, empowering product recommendations and finding the best alternative products.
[0089] In an embodiment, the price and margin sensitivity experimentation engine (218) may be used to:
(a) identify similar products using a product similarity module,
(b) for each of the merchant’s top products, recommend similar products with different margin/Maximum Retail Price (MRP),
(c) define the Expected Quantity, E(Q), for each top product. Using the expected Quantity, the margin sensitivity may be defined as mentioned below:
Margin Sensitivity = …. eq (5)
Where is actual quantity bought,
E(Q) is Estimated Quantity,
is Price of recommended products, and
is Price of top product.
(d) capture the merchant’s propensity to replace products as a response to changes in the margin and MRP. In yet another embodiment, one of the merchants may be elastic with respect to one product/category and inelastic with respect to another, and may be defined at a Merchant Category Level.
[0090] In an embodiment, the item similarity engine (220) facilitates to feed similar items or products to the price and margin sensitivity experimentation engine (218). To feed similar items to the price and margin sensitivity experimentation engine (218), a hybrid model may be created by combining term frequency-inverse document frequency (TF-IDF) vector-based similarity and Bidirectional Encoder Representations from Transformers (BERT) vector-based similarity. A person of ordinary skill in the art will understand that the TF-IDF is a statistical measure that evaluates relevancy of the products among multiple products. A BERT model is used for calculating semantic similarity between the products. Each of the TF-IDF vector-based similarity model and the BERT vector-based similarity model has their own advantages. For example, between items such as Brand A Biscuits and Brand B Cookies, the BERT model may capture semantic similarity between Biscuits and Cookies. On the contrary, in some edge cases, the TF-IDF output may always be noise free. The similarity of the items or the products may be computed individually based on both of the TF-IDF vector-based similarity model and the BERT vector-based similarity model. A final similarity score may be generated based on a weighted average of the scores obtained from the TF-IDF vector-based similarity model and the BERT vector-based similarity model.
[0091] Although FIG. 2 shows an exemplary block diagram (200) of the B2B recommendation system, in other embodiments, the B2B recommendation system (200) may include fewer components, different components, differently arranged components, or additional functional components than depicted in FIG. 1. Additionally, or alternatively, one or more components of the B2B recommendation system (200) may perform functions described as being performed by one or more other components of the B2B recommendation system (200).
[0092] FIG. 3 illustrates an exemplary architecture (300) of the B2B recommendation system, in accordance with an embodiment of the present disclosure. The disclosure caters to new e-commerce business where goods are sold to small merchants across the country. The disclosed system (e.g., 110 or 200) may make recommendations to the merchant whenever the merchant visits an e-commerce application.
[0093] With respect to FIG. 3, merchant transaction data (302) may be received as an input by a matrix factorization model (304), an exploration rate prediction model (306), a reorder prediction model (308), and a price and margin sensitivity experiments model (310). The merchant transaction data (302) may be used for new item prediction. The new item or product prediction is performed using the matrix factorization exploration rate prediction model (306), the reorder prediction model (308), and the price and margin sensitivity experiments model (310).
[0094] In an embodiment, for new item prediction, a collaborative filtering mechanism may be used to make recommendations for item prediction. The collaborative filtering mechanism may require learning from historical preferences of large groups of users and making predictions about the user’s likely preferences based on the learning. By way of an example, multiple mechanisms may make use of the matrix factorization model (304) for collaborative filtering. The disclosure employs a BPR algorithm that focuses on explicit likes and dislikes of the user and learns to predict the user preferences. However, the choice of the algorithm may not be limited to BPR and other similar algorithms may be employed.
[0095] In an embodiment, the exploration rate prediction model (306) may measure an exploration rate of the merchant, i.e., measure the distribution between a new product and old products at each transaction. Using the exploration rate prediction model (306), the prediction for reordering for the merchants (articles to entities) may be determined.
[0096] In an embodiment, the merchants may generally behave differently from the end users on B2C e-commerce applications since the merchant is largely driven by demand (from customers) and supply (from fulfilment centres) rather than individual preferences. The merchants typically often have a very substantial reordering habit, especially when their buying is looked at over large windows of time. This is also because when looked on overall basis, the end user’s behaviour does not change significantly for any merchant. Hence, the reorder prediction model (308) may be used to predict what products are likely to be ordered again by the merchants.
[0097] In yet another embodiment, the price and margin sensitivity experiments model (310) may be used to compute price and margin sensitivity of the merchant with respect to a product category.
[0098] Further, outputs from each of the matrix factorization model (304), the exploration rate prediction model (306), and the reorder prediction model (308) may be combined, and a weighted combination of the new and old prediction weights may be used to determine the exploration rate at (312). The determined exploration rate at (312) and the output from the price and margin sensitivity experiments model (310) may be received as input for post processing of recommendations at (314). The output from the price and margin sensitivity experiments model (310) may also be used to determine a sensitivity treatment and an out-of-stock replacement at (314). Thereafter, at (316), all the output from each of the models may be used to generate final recommendations.
[0099] Although FIG. 3 shows exemplary architecture (300) of the B2B recommendation system, in other embodiments, the architecture (300) may include fewer components, different components, differently arranged components, or additional functional components than depicted in FIG. 3. Additionally, or alternatively, one or more components of the network architecture (300) may perform functions described as being performed by one or more other components of the network architecture (300).
[00100] FIG. 4 illustrates an exemplary flow diagram (400) for generating recommendations using a matrix factorization engine, in accordance with an embodiment of the present disclosure.
[00101] With respect to FIG. 4, merchant transaction data (402) may be received and filtered for ‘N’ most recent transactions (404). Thereafter, a matrix factorization model (406) may factorize recent transactions into user factors (408) and item factors (410). Subsequently, the factors may be used to determine user item preferences (412).
[00102] FIG. 5 illustrates an exemplary flow diagram (500) for implementing a two-level reorder prediction engine, in accordance with an embodiment of the present disclosure.
[00103] With respect to FIG. 5, merchant transaction data (502) may be received as input by a two-level reorder prediction engine (e.g., 216) to create product forecast models at a city-product level at (504).
[00104] In an embodiment, the forecasting models (506) for say, product 1 to product N for city 1, a set of product – product 1, product 2,…., product N, at (508), may be used in a distribution module/engine (510) and a set of city 1 merchants, for example, merchant 1, merchant 2, …., merchant Z at (512).
[00105] In an embodiment, the forecasting models (506) may be received by one or more ranking modules (514). Further, for forecasting models for say, product 1 to product N for city N, the above-mentioned steps may be followed. The results of both the ranking modules (514) may be combined to determine a ranked reorder of recommendations at (516).
[00106] FIG. 6 illustrates an exemplary flow diagram (600) for implementing a price and margin sensitivity computation engine, in accordance with an embodiment of the present disclosure.
[00107] With respect to FIG. 6, merchant transaction data (602) may be shared with a product similarity model (604) and a merchant X (606).
[00108] In an embodiment, the data of the merchant X (606) may be divided into, for example, top products (608) for the merchant X (606) i.e., the products most bought for the merchant X (606) and estimated buying of top products (610) in next transactions.
[00109] In an embodiment, output from the product similarity model (604), and the determined top products (608) of the merchant X (606) may be used to determine similar item recommendations (614) for merchant X (606) having, for example, different MRP/margin. The merchant X (606) may buy items based on the recommendations (614).
[00110] In an embodiment, output from each of the estimated buying of top products (610) in next transaction for each top product (608), and from the buying recommendations (614) may be used to determine margin sensitivity of the item as illustrated.
[00111] FIG. 7 illustrates an exemplary flow diagram (700) for implementing an exploration rate prediction engine, in accordance with an embodiment of the present disclosure.
[00112] With respect to FIG. 7, merchant transaction data may be used along with a set of engineered features of the merchant. The engineered features of the merchant may include, but not limited to:
(a) features on platform (702) including an age on the platform and number of days since last purchase,
(b) amount spent based features (704) including an average amount spent per transaction and an amount spent in last transaction,
(c) item count based features (706) including an average number of unique/new items bought per transaction, and a number of unique items bought in last transaction,
(d) top products based features (708) including top products bought in last N transactions and top products bought in last one transaction, and
(e) auto regressive feature (710) including average distribution of new v/s old products in previous orders. Further, the merchant transaction data may be used along with the set of engineered features to extract a new item discovery rate model (712).
[00113] FIG. 8 illustrates an exemplary flow diagram (800) for implementing a product similarity engine, in accordance with an embodiment of the present disclosure.
[00114] With respect to FIG. 8, a product catalogue (802) may be received as input by the product similarity engine. In an embodiment, the product catalogue (802) may be divided into a TF-IDF Vectorizer (804) and a BERT Vectorizer (808). Using the TF-IDF Vectorizer (804), a matrix of N-Dimensional vectors (806) may be determined. Similarly, using the BERT Vectorizer (808), a matrix of N-Dimensional Vectors (810) may be determined.
[00115] In an embodiment, the matrix of N-Dimensional Vectors (806) and the matrix of N-Dimensional Vectors (810) may be combined to obtain vector similarity (812). Based on the vector similarity (812), each of a TF-IDF product similarity matrix (814) and a BERT product similarity matrix (816) may be obtained.
[00116] In an embodiment, the TF-IDF product similarity matrix (814) and the BERT product similarity matrix (816) may be used to calculate a weighted average (818) which may then be used to determine a product similarity matrix (820).
[00117] FIG. 9 illustrates an exemplary flow diagram (900) for implementing a distribution engine, in accordance with an embodiment of the present disclosure.
[00118] With respect to FIG. 9, both a city level forecast for a single product (902) and merchant transaction data (904) may be used to determine a distribution logic. In an embodiment, the distribution logic may include determining, at (906), (a) buying proportion of each merchant, and (b) an average of buying proportion over multiple periods for each merchant.
[00119] In an embodiment, the distribution logic may be used to determine distributed quantity over each of the merchants at (908).
[00120] As an advantage, improved quality of the recommendations may be provided to the merchants by predicting both new products that may be bought and also by informing what all previously bought products may be re-bought. This transforms the merchant’s shopping experience and may lead to minimizing user effort spent in manually searching for products on the platform or going through category pages.
[00121] FIG. 10 illustrates an exemplary flow diagram of a method (1000) for providing personalized recommendations to an end user, in accordance with an embodiment of the present disclosure.
[00122] With respect to FIG. 10, at block 1010, the method (1000) includes predicting user preference of a product by collecting historical preferences of a group of users. Further, at block 1020, the method (1000) includes predicting reordering of the product via a trained model, based on the user preference of the product, and predicting ordering of a new product.
[00123] At block 1030, the method (1000) includes determining an exploration rate between the new product and a reordered product based on a plurality of regression factors. Further, at block 1040, the method (1000) includes providing one or more personalized recommendations corresponding to the new product and the reordered product to the end user based on the exploration rate.
[00124] FIG. 11 illustrates an exemplary computer system (1100) in which or with which embodiments of the present disclosure may be implemented.
[00125] As shown in FIG. 11, the computer system (1100) may include an external storage device (1110), a bus (1120), a main memory (1130), a read only memory (1140), a mass storage device (1150), a communication port (1160), and a processor (1170).
[00126] A person skilled in the art will appreciate that the computer system (1100) may include more than one processor and communication ports. The processor (1170) may include various modules associated with embodiments of the present disclosure.
[00127] In an embodiment, the communication port (1160) may be any of an RS-232 port for use with a modem-based dialup connection, a 10/100 Ethernet port, a Gigabit or 10 Gigabit port using copper or fiber, a serial port, a parallel port, or other existing or future ports. The communication port (1160) may be chosen depending on a network, such a Local Area Network (LAN), Wide Area Network (WAN), or any network to which the computer system (1100) connects.
[00128] In an embodiment, the memory (1130) may be a Random Access Memory (RAM), or any other dynamic storage device commonly known in the art. The read-only memory (1140) may be any static storage device(s) e.g., but not limited to, a Programmable Read Only Memory (PROM) chips for storing static information e.g., start-up or Basic Input/Output system (BIOS) instructions for the processor (1170).
[00129] In an embodiment, the mass storage (1150) may be any current or future mass storage solution, which can be used to store information and/or instructions. Exemplary mass storage solutions include, but are not limited to, Parallel Advanced Technology Attachment (PATA) or Serial Advanced Technology Attachment (SATA) hard disk drives or solid-state drives (internal or external, e.g., having Universal Serial Bus (USB) and/or Firewire interfaces), one or more optical discs, Redundant Array of Independent Disks (RAID) storage, e.g. an array of disks (e.g., SATA arrays).
[00130] In an embodiment, the bus (1120) communicatively couples the processor(s) (1170) with the other memory, storage, and communication blocks. The bus (1120) may be, e.g., a Peripheral Component Interconnect (PCI) / PCI Extended (PCI-X) bus, Small Computer System Interface (SCSI), Universal Serial Bus (USB) or the like, for connecting expansion cards, drives, and other subsystems as well as other buses, such a front side bus (FSB), which connects the processor (1170) to computer system (1100).
[00131] Optionally, operator and administrative interfaces, e.g., a display, keyboard, joystick, and a cursor control device, may also be coupled to the bus (1120) to support direct operator interaction with the computer system (1100). Other operator and administrative interfaces may be provided through network connections connected through the communication port (1160). Components described above are meant only to exemplify various possibilities. In no way should the aforementioned exemplary computer system (1100) limit the scope of the present disclosure.
[00132] While the foregoing describes various embodiments of the invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof. The scope of the invention is determined by the claims that follow. The invention is not limited to the described embodiments, versions or examples, which are included to enable a person having ordinary skill in the art to make and use the invention when combined with information and knowledge available to the person having ordinary skill in the art.

ADVANTAGES OF THE PRESENT DISCLOSURE
[00133] The present disclosure provides personalized product recommendations to merchants and/or distributors.
[00134] The present disclosure provides product recommendations to the merchants and/or distributors in a Business-to-Business (B2B) scenario.
[00135] The present disclosure provides product recommendations while improving quality of the recommendations made by predicting both new products that may be bought and also the products that were previously bought by users, thereby leading to significant uplifting of the merchant’s buying experience.
[00136] The present disclosure transforms shopping experience of the merchants.
[00137] The present disclosure provides the recommendations to ease or reduce manual searching of the products and minimize user effort.
[00138] The present disclosure determines price/margin sensitivity that enables to tackle out of stock situations as well as provides additional information to improve recommendations of the products.
,CLAIMS:1. A system (200) for providing personalized recommendations, the system (200) comprising:
one or more processors (202); and
a memory (204) operatively coupled to the one or more processors (202), wherein the memory (204) comprises processor-executable instructions, which on execution, cause the one or more processors (202) to:
predict user preference of at least one product by collecting historical preferences of a group of users;
predict, via a trained model, reordering of the at least one product and ordering of a new product based on the user preference of the at least one product;
determine an exploration rate between the new product and at least one reordered product based on a plurality of regression factors; and
provide one or more personalized recommendations corresponding to the new product and the at least one reordered product to an end user based on the exploration rate.

2. The system (200) as claimed in claim 1, wherein the memory (204) comprises processor-executable instructions, which on execution, cause the one or more processors (202) to determine price and margin sensitivity of a distributor with respect to a category of the at least one product.

3. The system (200) as claimed in claim 1, wherein the memory (204) comprises processor-executable instructions, which on execution, cause the one or more processors (202) to determine a change in price and margin of the at least one reordered product, and replace the at least one reordered product with an alternative product based on the determined change.
4. The system (200) as claimed in claim 1, wherein the one or more processors (202) are to predict the reordering of the at least one product and the ordering of the new product by being configured to:
train a respective model for each combination of a unique area and a unique product;
determine, via the trained model, whether the at least one product is ordered multiple times or the at least one product is ordered for a first time based on the historical preferences of the group of users;
predict the reordering of the at least one product in response to determining that the at least one product is ordered multiple times; and
predict the ordering of the new product in response to determining that the at least one product is ordered for the first time.

5. The system (200) as claimed in claim 1, wherein the memory (204) comprises processor-executable instructions, which on execution, cause the one or more processors (202) to optimize a balance between the reordering of the at least one product and the ordering of the new product based on the exploration rate.

6. The system (200) as claimed in claim 1, wherein the memory (204) comprises processor-executable instructions, which on execution, cause the one or more processors (202) to:
identify one or more similar products by creating a hybrid model, wherein the hybrid model is created by combining a Term Frequency-Inverse Document Frequency (TF-IDF) vector-based similarity and a Bidirectional Encoder Representation from Transformers (BERT) vector-based similarity; and
provide the one or more personalized recommendations corresponding to the one or more similar products to the end user using the hybrid model.

7. The system (200) as claimed in claim 1, wherein the memory (204) comprises processor-executable instructions, which on execution, cause the one or more processors (202) to provide the one or more personalized recommendations corresponding to at least one alternative product based on the at least one product ordered by the end user being out of stock.

8. The system (200) as claimed in claim 1, wherein the plurality of regression factors comprises at least one of: an age on a distribution platform, number of days after last purchase of the at least one product, an average amount spent per transaction, an amount spent in last transaction, and an average number of unique products bought per transaction.

9. A method for providing personalized recommendations, the method comprising:
predicting, by a processor (202) associated with a system (200), user preference of at least one product by collecting historical preferences of a group of users;
predicting, by the processor (202) via a trained model, reordering of the at least one product and ordering of a new product based on the user preference of the at least one product;
determining, by the processor (202), an exploration rate between the new product and at least one reordered product based on a plurality of regression factors; and
providing, by the processor (202), one or more personalized recommendations corresponding to the new product and the at least one reordered product to an end user based on the exploration rate.

10. The method as claimed in claim 9, comprising determining, by the processor (202), price and margin sensitivity of a distributor with respect to a category of the at least one product.

11. The method as claimed in claim 9, comprising determining, by the processor (202), a change in price and margin of the at least one reordered product, and replacing, by the processor (202), the at least one reordered product with an alternative product based on the determined change.

12. The method as claimed in claim 9, wherein predicting, by the processor (202) via the trained model, the reordering of the at least one product and the ordering of the new product comprises:
training, by the processor (202), a respective model for each combination of a unique area and a unique product;
determining, by the processor (202) via the trained model, whether the at least one product is ordered multiple times or the at least one product is ordered for a first time based on the collected historical preferences of the group of users;
predicting, by the processor (202), the reordering of the at least one product in response to determining that the at least one product is ordered multiple times; and
predicting, by the processor (202), the ordering of the new product in response to determining that the at least one product is ordered for the first time.

13. The method as claimed in claim 9, comprising optimizing, by the processor (202), a balance between the reordering of the at least one product and the ordering of the new product based on the exploration rate.

14. The method as claimed in claim 9, comprising:
identifying, by the processor (202), one or more similar products by creating a hybrid model, wherein the hybrid model is created by combining a Term Frequency-Inverse Document Frequency (TF-IDF) vector-based similarity and a Bidirectional Encoder Representation from Transformers (BERT) vector-based similarity; and
providing, by the processor (202), the one or more personalized recommendations corresponding to the one or more similar products to the end user using the hybrid model.

15. The method as claimed in claim 9, comprising:
determining, by the processor (202), that the at least one product ordered by the end user is out of stock; and
providing, by the processor (202), the one or more personalized recommendations corresponding to at least one alternative product in response to determining that the at least one product ordered by the end user is out of stock.

16. The method as claimed in claim 9, wherein the plurality of regression factors comprises at least one of: an age on a distribution platform, number of days after last purchase of the at least one product, an average amount spent per transaction, an amount spent in last transaction, and an average number of unique products bought per transaction.

17. A user equipment (104), comprising:
one or more processors; and
a memory operatively coupled to the one or more processors, wherein the memory comprises processor-executable instructions, which on execution, cause the one or more processors to:
receive one or more personalized recommendations from a system (200) via a wireless network;
wherein the one or more processors are communicatively coupled with the system (200), and wherein the system (200) is configured to:
predict user preference of at least one product by collecting historical preferences of a group of users;
predict, via a trained model, reordering of the at least one product and ordering of a new product based on the user preference of the at least one product;
determine an exploration rate between the new product and at least one reordered product based on a plurality of regression factors; and
provide the one or more personalized recommendations corresponding to the new product and the at least one reordered product to an end user associated with the user equipment (104) based on the exploration rate.