DESC:FIELD OF THE INVENTION

The present invention is in the field of information technology and ecommerce, and more particularly facilitating the matching or coincidence of desires and wishes of parties be prospective buyers and sellers of any products and services, offered for sale, exchange or bartering.

PROBLEM AND BACKGROUND BEHIND THE INVENTION:

Relevant matching of somebody’s desire with somebody else’s wish has been an active area of prolonged research. Thus there exists a need to develop this area and the same is being achieved by creating computational methods and system to efficiently tackle the problem of matching the coincidence of desire between two parties. The idea behind this invention is to develop or design a system which will efficiently deal the prevailing problem of relevant matches between desires and wishes of two parties, buyers and sellers, in a sale or exchange or bartering system. The problem which exists today is non- availability of any system or portal for proper matching of coincidence of desires and wishes of prospective buyers and sellers having something to offer be as a product or service, based only on any text analysis. This process has been reconsidered by adopting a novel machine learning model using Name Entity recognition (NER) property of text.

SUMMARY OF THE INVENTION

According to a preferred embodiment of the invention, a system for coincidence of desires between prospective buyers and sellers of any offers, comprising a plurality of data collection devices, each connected to at least one packet- based data network and adapted to collect data pertaining to a plurality of prospective buyers or sellers of complex offers, a summary data generator software module operating on a server computer and connected via a data network to a database, an attribute index generator software module operating on a server computer and connected via a data network to the database, a categorization software module operating on a server computer and connected via a data network to the database, a buyer analysis engine software module operating on a server computer and connected via a data network to the database, an analysis engine software module operating on a server computer and connected via a data network to the database, and a matching engine software module operating on a server computer and connected via a data network to the database is disclosed.

According to the embodiment, data collected by the data collection devices is stored in the database and is used by the summary data generator software module to generate a plurality of summary data elements pertaining to the desires of prospective buyer and seller of an offer, and the plurality of summary data elements is stored in the database and used by the attribute index generator software module to generate attribute indices each based on at least two summary data elements, and at least some data collected by the data collection devices is used by the buyer analysis engine software module to determine at least a probability that a buyer will buy a specific offer or a seller will be successful in selling a specific offer, and the matching engine software module is used to determine an optimal matching of coincidence of desires of prospective buyers and sellers on offers based on scores as generated and a likelihood to buy for each prospective buyers. Further location matching of prospective buyers and sellers are also carried out in the process based on scores as generated for location matching computed by Havershine formula, which calculates the shortest distance between latitude and longitude pairs.

Thus this invention will help to achieve to develop or design a system which will efficiently deal the prevailing problem of relevant matches between desires and wishes of two parties, buyers and sellers, in a sale, exchange or bartering system.

BRIEF DESCRIPTION OF INVENTION

The invention provides, in one embodiment, a system for matching coincidence of desires of prospective buyers and sellers of any products and services. In the process various data collection modules are used to collect a wide range of data about desires of prospective buyers and sellers.

Data sometimes can be collected directly from users via a plurality of web pages for example when a prospective buyer fills out a survey discussing his / her prospective needs in a particular area such as sustainability products and services or by any other applicable methods.

Generally data collected by a plurality of data collectors is sent to and stored in data storage module.

Data store is embedded in a preferred embodiment a relational database management system (RDBMS), as available from reputed providers.

According to the preferred embodiment, raw data is sent from data storage module to summary data generator, a software module operating on one or more general-purpose computers.

Summary data generator applies rules to raw data obtained from data storage module to generate a large number of summary data elements, which are then passed to data storage module for storage and later use.

Summary data is obtained from data storage module by applicable and available attribute index generator , a software module that applies a set of configurable weighted relational process to various subsets of summary data to produce a plurality of high-level indices that correspond to a company’s or

organization’s level of need for a particular category of sustainability product or service.

The process also involves use of a user interface which provides a means for human users to interact with various software modules of the invention.

For instance, matching engine could be used by a user (via user interface) to create a list of leads for products and services as available. Then, using a matching process, as above, desires of prospective buyers are matched with the most appropriate desires of products and services being offered, and is returned to the user (via user interface), generally as a ranked list of prospects with assessments of the prospective buyers and sellers matching the coincidence of their desires. Matching is done, according to a preferred embodiment, using the said optimized method, the objective being to optimize the degree of match between desires of buyers and proposed products or services.

Considering now the case when a user is a seller, the user completes (or updates) a seller profile in the process step, just like a buyer. This step is composed of several sub-steps that may be performed in any order.

The matching method has been designed to make the bartering concept both in product and service more efficient and effective. In our proposed methodology, the user’s provides information like the description, item/service category, user’s desire description, desiring item/service tag are taken into account along with user’s id as input. Then the matched listings are retrieved by using these input information.

listingType: It says whether a “product” or “service” is being listed/posted.

listingDescription: It describes what the user is listing. It contains the brand, colour, item name while a product is being listed and the relevant description is put in case of service while a service is listinged. It’s a collection of typical English sentences.

listingTag: listingTag contains the category selected by the user from our category taxonomy.

listingName: It is a combination of listing category and mandatory attributes.

TransactionMode: In our scenario it is barter.

desireType: It denotes what a user desires among “product”, “service” and “Anything” in return while bartering product/service.

desire: It is the description of the user’s desire that may consist of a number of typical English sentences.

desireTag: desireTag contains the category selected by the user from our category taxonomy.

UserId: It is a unique identifier of the user.

I. Data Cleaning:
The above-described textual inputs of listing and desire are grouped and concatenated to form the modified description (d1) and modified desire (d2).

The stop words from the text are removed in cleanText process and the lemmatized form of the processed text is passed for the next cleaning phase. In the next cleaning phase all full stops(.), double quotations(“), single quotations(‘), and commas(,) are removed when they appear in any non-numeric word token. All adverbs are removed and all verbs that do not follow the specified patterns like [{Noun-Verb},
{Verb-Noun}, {Punctuation-Verb}, {Verb-Punctuation}] are also removed from the processed text. One more thing has been incorporated in the processing is considering the word which has two parts text1 and text2 joined by hyphen(-) like text1-text2.

II. Indexing:
The powerful and advantageous application of Elastic Search has been incorporated in our listing indexing. Method:1 is applied to all the existing published listings to generate the search field set cleanDescriptionInformation(P), cleanDesireInformation(Q) and all listing related information stored in the listing entity (G).

P: = { ?? | i, i=1,2,3……N} eq:(1)

??
1
Q: = {??

| i, i=1,2,3……N} eq:(2)

2
G:= {(??????????????) | i, i=1,2,3……N} eq:(3)
??
where N= total number of published listing at any given time.

III. Query :

Query in our document is referred to as a new listing uploaded in our platform or any existing listing for which the matching method will run. Based on the query, we are fetching the set of listings for further processing.

This is explained in FIG 4 the initial result fetching process from Elastic Search engine.
For all the existing published listings, cleanDescriptionInformation (P) and cleanDesireInformation (Q) are generated and indexed in ElasticSearch engine. Now, when a query listing comes, we generate modified description (d1) and modified desire (d2) to fetch the required set of listings. There are two parallel processes executed during listing matching which is shown in Fig. 2. In one case, modified description (d1) & pre-indexed cleanDesireInformation (Q) are involved and in another case modified desire (d2) & pre-indexed cleanDescriptionInformation (P) are involved.

In this process searchMatchResult (described in Method:2) is called first with the parameters (Q, d1, IdentityId, listingType/desireType) as shown in eq(4) and the set of listing (matchDescriptionListings) is generated based on the text match score.

matchListings= searchMatchResult(Q, d1, IdentityId, listingType) eq:(4)

In this method it is restricted that if the user is a “”then that will participate in match score generation whose desireType is “” only. Similarly, if “service” is then in matching no will participate whose desire is a “”. The output of eq. 4 is
normalized (See Method:2a) and passed as the matched result when desireType=”Anything”.

f the desireType is not “Anything” then the process starts the 2nd pipeline with the modifiedDesire (d2). In this process, searchMatchResult is called with new the parameters (P, d2, IdentityId, listingType/desireType) and the set of listing (matchDesireListings) is generated based on the text match score.

Finally, the two sets “matchDescriptionListings” and “matchDesire” are merged by union operation as shown in eq:(5).

unionMap=union(matchDescriptionListings,matchDesire) eq:(5)

I. Matched Score aggregation:

From the unionMap created in the above section, we extracted the unique IDs of the listings to aggregate the match score as shown in Fig. 5.

The scores of in unionMap are aggregated by their participation in description matching and desire matching. As discussed in Method:3, if a listing of unionMap is present in both “matchDescription” and “matchDesire” then the two scores store the value into corresponding “matchScore”. If the is present in either of the two sets, then it’s
In Tag score generation as explained in Method:4 the process generates two different scores “tagScore” and “synoTagScore”.
In “tagScore'' generation process two tag similarity scores (s1 and s2) are calculated based on the natural language toolkit corpus. s1 is the tag similarity score between “Tag” and “desireTag '' of the in unionMap. Fig 6.

Similarly, s2 is obtained by evaluating the similarity score between “desireTag” and the “Tag” of the in unionMap. Finally, the tag score is by taking the average of s1 and s2.
In synoTagScore generation, we rely on the value of two parameters a and ß. Here value is assigned to a on the basis of following conditions,

Cond1: If any word in esireTag is present in Tag of in unionMap then a=0.5.
Cond2: If Cond1 fails and any synonym generated by the natural language toolkit corpus of is present in of the in unionMap then a =0.42.
Cond3: If Cond1 and Cond2 fail then a=0

Similarly, ß is assigned a value by checking the presence of word of qTag and synonym of Tag in desireTag in unionMap.

The obtained values of a and ß are added to get the synoTagScore which is summed with tagScore to find out the combinedScore. The “semiFinalScore” is calculated by putting some empirical weightages on both the obtained combined score and ’s matchScore.

Function f(dist) calculates the distance between the and the participated in the matching. The less distance score is considered high and the far distance score is considered less. Similarly, the function f(price) calculates the price difference between the ’s price and the price of the other matched listings.

Lastly, the matchScore of the is set to finalScore with the value of f(dist) and f(price). In this way the matchScore of all in unionMap is aggregated and the final unionMap is sorted as per its ’ matchScore value. At last the desired are retrieved and these are kept in resultMap as the final match result and is displayed in the website based on user interaction

When a plurality of seekers and sellers have created profiles, and incrementally enriched those profiles in subsequent iterations through applicable steps respectively, and when a plurality of seeker portfolios and needs and a plurality of seller solution portfolios have been created, then in next step matching module is applied by matching engine to determine an optimal or near-optimal, or at least a desirable coincidence of desires list of proposed buyer/seller pairings, and particular solutions associated therewith.

While the disclosure and features have been described in terms of exemplary embodiments, those skilled in the art will recognize that the disclosure can be practiced with modifications in the spirit and scope of the applicable claims.

These examples given above are merely illustrative and are not meant to be an exhaustive list of all possible designs, embodiments, applications, or modifications of the disclosure. Even though the invention is aimed at providing a system for matching the coincidence of desire and wishes of prospective buyers and sellers, it can be used and modified to match other similar attributes of buyers and sellers too.

Various embodiments of the present invention / technology provide certain advantages. Not all embodiments of the invention / technology share the same advantages and those that do may not share them under all circumstances.

Moreover the above description and explanation of the invention / technology are not an exhaustive listing of the forms pertaining to a system for matching the coincidence of desire and wishes of prospective buyers and sellers, in accordance with the invention might take, rather they serve as the exemplary basics as presently understood.

,CLAIMS:WE CLAIM:

1. A method for matching the coincidence of desire between two parties mainly between prospective buyers and sellers of any offers comprising:

(i) providing a buyer database for storing data entered by a plurality of buyers registered on a buyer-seller matching website, said buyer database comprising a plurality of data structures, each data structure of the plurality being assigned to store the data entered by a single buyer of the plurality of registered buyers, the number of data structures equaling the number of registered buyers

(ii) prompting a first user to log onto the buyer-seller matching website as a buyer for matching the coincidence of desire between two parties and enter personal data;

(iii) prompting the buyer to enter one or more coincidence of desire search attribute descriptors and to assign a numerical importance score to each of the one or more entered coincidence of desire attribute descriptors;

(iv) creating data structures in the buyer database for storing the personal data, search attribute descriptor data and numerical importance score data entered by the buyer;

(v) storing each of said one or more search attribute descriptors and associated numerical importance score values assigned thereto that are entered by the buyer in the created data structures located in the buyer database, wherein created data structures hold the one or more search attributes / coincidence of desire and the associated numerical importance scores entered by the buyer the created data structures are added to the plurality of buyer data structures stored in the buyer database;

(vi) providing a seller database for storing data entered by a plurality of sellers registered on a buyer-seller matching website, said seller database comprising a plurality of data structures, each data structure of the plurality being assigned to store the data entered by a single seller of the plurality of registered sellers;

(vii) prompting a second user to log onto the buyer-seller matching website as a seller searching for a coincidence of desire in any product as offered for sale;

(viii) prompting the seller to enter one or more coincidence of desire attribute descriptors;

(ix) creating one or more data structures in the seller database for storing the personal data descriptors and coincidence of desire attribute descriptors data entered by the seller;

(x) storing each of said one or more such coincidence of desire attribute descriptors that are entered by the seller in the one or more created data structures created in the seller database, wherein the created data structures that hold the one or more coincidence of desire attributes entered by the seller are added to the plurality of seller data structures stored in the seller database;

(xi) comparing each of the coincidence of desire search attribute descriptors entered by the plurality of buyers to the coincidence of desire attribute descriptors entered by the seller;

(xii) choosing among the plurality of buyers those buyers that have at least one search attribute matching a coincidence of desire attribute entered by the seller for gathering a group of buyers to present to the seller of the offered product;

2. The method of claim 1, wherein the step of prompting the buyer to enter one or more coincidence of desire search attributes and to assign a numerical importance score to each of the one or more entered coincidence of desire attributes is enabled by a graphical interface on the buyer-seller matching coincidence of desire website, said graphical interface comprising a plurality of data entry windows, each data entry window corresponding to a unique coincidence of desire feature chosen from a plurality of coincidence of desire features.