CN114155049B - Method and device for determining target object - Google Patents
Method and device for determining target object Download PDFInfo
- Publication number
- CN114155049B CN114155049B CN202210117160.9A CN202210117160A CN114155049B CN 114155049 B CN114155049 B CN 114155049B CN 202210117160 A CN202210117160 A CN 202210117160A CN 114155049 B CN114155049 B CN 114155049B
- Authority
- CN
- China
- Prior art keywords
- supply
- objects
- demand
- word vector
- determining
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The invention relates to a method and a device for determining a target object, wherein the method comprises the following steps: determining embedded representations of a plurality of objects according to a predetermined supply-demand relationship diagram; determining target objects having a similar supply chain as the current object based on the embedded representations of the respective objects; wherein the plurality of objects includes: the current object and the target object; the supply and demand relation graph takes an object with a supply and demand relation as a node and the supply and demand relation as an edge, and the object with the supply and demand relation is determined by the supply and demand behavior data of the objects. The method can accurately determine the target object with a supply chain similar to the current object based on the supply-demand relation graph, and can efficiently obtain the result by performing operation based on the embedded representation of the object.
Description
Technical Field
The present invention relates to the field of computer technologies, and in particular, to a method and an apparatus for determining a target object.
Background
In actual production and operation activities, upstream objects and downstream objects form a supply chain by virtue of supply-demand relationships. For example, company a sells chemical raw materials to company B, and company B sells organic intermediates to company C, then company a, company B, and company C form a supply chain. For company B, company A is its upstream company and company C is its downstream company.
For two objects which are competitors to each other, the supply chains of the two objects are similar, and the objects with competitive relationship can be further discovered by determining the objects with similar supply chains. Therefore, how to determine the target object with a supply chain similar to the current object becomes a problem which is more concerned by the technical staff.
Disclosure of Invention
The invention aims to provide a method and a device for determining a target object so as to determine objects with similar supply chains.
In a first aspect, an embodiment of the present invention provides a method for determining a target object, where the method includes:
determining embedded representations of a plurality of objects according to a predetermined supply-demand relationship diagram;
determining target objects having a similar supply chain as the current object based on the embedded representations of the respective objects;
wherein the plurality of objects includes: the current object and the target object; the supply and demand relation graph takes an object with a supply and demand relation as a node and the supply and demand relation as an edge, and the object with the supply and demand relation is determined by the supply and demand behavior data of the objects.
Alternatively,
determining an embedded representation of a plurality of objects according to a predetermined supply-demand relationship graph, comprising:
generating a plurality of supply and demand paths according to the supply and demand relationship diagram; wherein the supply and demand path comprises a plurality of objects;
generating embedded representations of the plurality of objects according to the plurality of supply and demand paths.
Alternatively,
generating a plurality of supply and demand paths according to the supply and demand relationship diagram, wherein the generating comprises:
for each node in the supply-demand relationship graph:
and randomly selecting edges connected with the nodes by taking the nodes as a starting point, and determining the nodes in the supply and demand paths according to the selected edges to obtain the supply and demand paths passing through the nodes of the first number.
Alternatively,
the supply and demand behavior data comprises: a transaction amount;
generating a plurality of supply and demand paths according to the supply and demand relationship diagram, wherein the generating comprises:
for each node in the supply-demand relationship graph:
selecting edges connected with the nodes according to the weights of the edges by taking the nodes as starting points, and determining the nodes in the supply and demand paths according to the selected edges to obtain the supply and demand paths which are routed to the second number of nodes; wherein the weight of the edge is determined by the transaction amount between the nodes.
Alternatively,
generating, by the computing device, embedded representations of the plurality of objects according to the plurality of supply and demand paths, including:
and inputting the multiple supply and demand paths into a word vector model to obtain the embedded representation of the multiple objects.
Alternatively,
said determining a target object having a similar supply chain as the current object based on the embedded representation of each said object, comprising:
determining a similarity of the embedded representation of the current object to the embedded representations of the respective other objects;
selecting a third number of objects from a plurality of other objects as the target objects according to the sequence of similarity of the embedded representation of the current object from large to small; wherein the similarity between the embedded representation of the target object and the embedded representation of the current object is used for representing the similarity between the supply chain of the target object and the supply chain of the current object.
Alternatively,
said determining a target object having a similar supply chain as the current object based on the embedded representation of each said object, comprising:
adding the current object to a search set and recall set;
according to the first-in first-out sequence, first objects ranked at the first position are obtained from the search set, and a plurality of second objects with similarity larger than a first threshold value with the first objects are searched in the plurality of objects;
for each of the second objects: terminating the current flow if the second object is present in the recall set; adding the second object to the search set and the recall set if the second object is not present in the recall set;
deleting the first object from the search set;
determining whether the number of second objects in the recall set is greater than a fourth number or whether the search set is empty, and if so, determining that the second objects in the recall set are the target objects; otherwise, executing the first-in first-out sequence, acquiring a first object arranged at the first position from the search set, and searching a plurality of second objects with the similarity to the first object being greater than a first threshold value;
wherein the similarity between the embedded representation of the target object and the embedded representation of the current object is used for representing the similarity between the supply chain of the target object and the supply chain of the current object.
In a second aspect, an embodiment of the present invention provides an apparatus for determining a target object, including:
an embedded representation determining module configured to determine embedded representations of the plurality of objects according to a predetermined supply-demand relationship diagram;
an object determination module configured to determine a target object having a similar supply chain as a current object based on the embedded representation of each of the objects; wherein the plurality of objects includes: the current object and the target object; the supply and demand relation graph takes an object with a supply and demand relation as a node and the supply and demand relation as an edge, and the object with the supply and demand relation is determined by the supply and demand behavior data of the objects.
In a third aspect, an embodiment of the present invention provides a non-transitory computer-readable storage medium, on which a computer program is stored, and the computer program, when executed by a processor, implements the steps of the method according to any one of the above embodiments.
In a fourth aspect, an embodiment of the present invention provides an electronic device, including:
a memory having a computer program stored thereon;
a processor for executing the computer program in the memory to implement the steps of the method of any of the above embodiments.
By adopting the technical scheme, the following technical effects can be at least achieved: the embodiment of the invention constructs a supply and demand relation graph based on the supply and demand relation among the objects, and extracts the embedded representation of the objects from the supply and demand relation graph. Because the embedded representation of the object can reflect the supply and demand characteristics of the supply chain in which the object is positioned, the target object with the supply chain similar to the current object can be accurately determined based on the embedded representation of the object. In addition, the method performs calculation based on the embedded representation of the object, and can obtain the result more efficiently.
Additional features and advantages of the invention will be set forth in the detailed description which follows.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention. In the drawings:
FIG. 1 is a flow diagram illustrating a method of determining a target object in accordance with one embodiment of the present invention;
FIG. 2 is a schematic diagram of a supply and demand relationship diagram according to an embodiment of the present invention;
FIG. 3 is a flow diagram illustrating a method for determining businesses with similar supply chains, in accordance with an embodiment of the present invention;
FIG. 4 is a schematic structural diagram illustrating an apparatus for determining a target object according to an embodiment of the present invention;
fig. 5 is a block diagram illustrating an electronic device in accordance with an exemplary embodiment of the present invention.
Detailed Description
The following detailed description of embodiments of the invention refers to the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating the present invention, are given by way of illustration and explanation only, not limitation.
It should be understood that the various steps recited in method embodiments of the present invention may be performed in a different order, and/or performed in parallel. Moreover, method embodiments may include additional steps and/or omit performing the illustrated steps. The scope of the invention is not limited in this respect. The term "include" and variations thereof as used herein are open-ended, i.e., "including but not limited to". The term "based on" is "based, at least in part, on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Relevant definitions for other terms will be given in the following description.
It should be noted that the terms "first", "second", and the like in the present invention are only used for distinguishing different devices, modules or units, and are not used for limiting the order or interdependence relationship of the functions performed by the devices, modules or units. It is noted that references to "a", "an", and "the" modifications in the present invention are intended to be illustrative rather than limiting, and that those skilled in the art will recognize that reference to "one or more" unless the context clearly dictates otherwise.
In a practical application scenario, an upstream object and a downstream object form a supply chain by virtue of supply-demand relationships. The object may be an enterprise, an individual industrial business, or the like. For two objects that are competitors to each other, they tend to have a more similar supply chain. Therefore, in an actual application scenario, objects with competitive relationships can be mined by determining objects with similar supply chains, and then corresponding marketing strategies can be formulated for competitors.
In view of this, the following embodiments will explain in detail the determination method of objects having a similarity supply chain.
As shown in fig. 1, an embodiment of the present invention provides a method for determining a target object, including:
step 101: the embedded representations of the plurality of objects are determined from a predetermined supply-demand relationship graph.
The supply and demand relation graph is formed by objects with supply and demand relations, wherein the objects are nodes, and the supply and demand relations among the objects are edges. As shown in fig. 2, the node includes: object 1, object 2, object 3 and object 4, object 2 having an on-demand relationship with object 1, object 3 and object 4 respectively, and object 1 having an on-demand relationship with object 4. The supply and demand relationship between the nodes can be unidirectional or bidirectional. Taking object 1 and object 2 as examples, in case 1, object 1 purchases chemical raw material from object 2, then object 1 is the demand side, and object 2 is the supply side; in case 2, object 2 purchases a chemical product from object 1, then object 2 is a demand side and object 1 is a supply side. In the embodiment of the present invention, only case 1 or case 2 may exist, and case 1 and case 2 may also exist at the same time.
The embedded representation of the objects, i.e. the word vectors of the respective objects, is extracted from the supply-demand relationship graph. For example, the embedded representations of object 1 through object 4 are extracted from FIG. 2. The embedded representation of the object can reflect the supply and demand characteristics of the supply chain in which the object is located.
Step 102: from the embedded representations of the respective objects, a target object having a similar supply chain as the current object is determined.
Wherein the plurality of objects includes: the current object and the target object. It should be noted that, other objects besides the current object and the target object may also be included in the plurality of objects.
For example, the embedded representations of the object 1 to the object 4 are determined according to the supply-demand relationship diagram, and if the current object is the object 1, the target object is determined among the object 2 to the object 4 according to the embedded representations of the object 1 to the object 4. Similarly, if the current object is object 2, a target object is determined among object 1, object 3, and object 4.
In the embodiment of the present invention, the target object may be determined according to a preset second threshold, specifically, the similarity between the supply chain of the current object and the supply chain of the target object is greater than the preset second threshold, that is, if the similarity between the supply chain of the object and the supply chain of the current object is greater than the second threshold, the object is determined to be the target object. The second threshold may be set by the user according to actual business requirements, for example, if the user only focuses on target objects with a similarity greater than 90% to the supply chain of the current object, the second threshold may be preset to 90%.
In the embodiment of the present invention, the target objects may also be determined according to a preset number, such as the third number and the fourth number in the following embodiments. For example, if the number of target objects that the user wants to acquire is 3, the objects ranked in the top 3 are the target objects in the order of the similarity of the supply chain from high to low.
The embodiment of the invention constructs a supply and demand relation graph based on the supply and demand relation among the objects, and extracts the embedded representation of the objects from the supply and demand relation graph. Because the embedded representation of the object can reflect the supply and demand characteristics of the supply chain where the object is located, the target object with the supply chain similar to the current object can be accurately determined based on the embedded representation of the object. In addition, the method can perform calculation based on the embedded representation of the object, and can determine the target object more efficiently. The target object may have a competitive relationship with the current object and may also have a potential cooperative relationship with the current object, for example, enterprise 1 selling projectors, enterprise 2 selling projectors and enterprise 3 selling desks all have a supply relationship with the training institution 4, enterprise 1, enterprise 2 and enterprise 3 have similar supply chains, enterprise 2 is a competitor of enterprise 1, enterprise 3 is a potential cooperative object of enterprise 1, and enterprise 3 may cooperate with enterprise 1 to provide products for the training institution 4.
In one embodiment of the invention, the method further comprises:
acquiring supply and demand behavior data of each object;
determining objects with supply and demand relations according to the supply and demand behavior data of each object;
and generating a supply and demand relation graph by taking the object with the supply and demand relation as a node and the supply and demand relation as an edge.
In the embodiment of the present invention, the supply and demand behavior data may include any one or more of the following: bid data, contract data, and invoice data. The bidding data may include any one or more of the following: bidding documents, bid-winning announcements, bid-winning notices, and the like. The tenderer and the middle tenderer can be obtained from the tendering and bidding data, the first party and the second party can be obtained from the contract data, and the buyer and the seller can be obtained from the invoice data. The tenderer, the buyer and the winning party, the selling party and the selling party are suppliers, and the suppliers and the demanders have supply-demand relations.
The supply and demand behavior data may include transaction products, transaction amounts, and the like in addition to the demander and the supplier. For example, the transaction amount includes: medium bid price, contract amount, and invoice amount.
According to the embodiment of the invention, the supply and demand relation between the objects is determined through the supply and demand behavior data, and then the connection relation between the nodes is determined, so that a supply and demand relation graph is obtained. According to the supply-demand relationship diagram, the characteristics of the supply chain in which the object is located, such as the object contained in the supply chain, can be determined. The supply and demand relation diagram not only considers the supply behavior (such as selling) of the object, but also considers the demand behavior (such as purchasing) of the object, and the target object with a similar supply chain with the current object can be accurately determined based on the supply and demand relation diagram.
In one embodiment of the invention, determining an embedded representation of a plurality of objects according to a predetermined supply-demand relationship graph comprises:
generating a plurality of supply and demand paths according to the supply and demand relationship diagram; wherein, the supply and demand path comprises a plurality of objects;
an embedded representation of the plurality of objects is generated based on the plurality of supply and demand paths.
One supply and demand path represents one supply chain, and the number of objects included in each supply and demand path can be preset according to the demand, for example, the number of objects included in the preset supply and demand path does not exceed 3. Taking fig. 2 as an example, the generated supply and demand paths include, but are not limited to: object 1-object 2-object 3, object 1-object 4-object 2, object 2-object 3. If the number of objects included in the preset supply and demand path is 4, the generated supply and demand path includes, but is not limited to: object 1-object 4-object 2-object 3, object 3-object 2-object 1-object 4.
In order to determine the target object more quickly, the embodiment of the invention converts the supply and demand paths extracted from the supply and demand relationship diagram into the embedded representation, so that the final target object is obtained through vector operation subsequently.
In practical application scenarios, there are at least two ways to determine supply and demand paths.
Mode 1:
generating a plurality of supply and demand paths according to the supply and demand relationship diagram, wherein the method comprises the following steps:
for each node in the supply-demand relationship graph:
and randomly selecting edges connected with the nodes by taking the nodes as a starting point, and determining the nodes in the supply and demand paths according to the selected edges to obtain the supply and demand paths passing through the first number of nodes.
The first number is the number of nodes included in a preset supply and demand path. In the embodiment of the invention, when a plurality of edges connected with the nodes exist, the probability of each edge being selected is the same. Taking fig. 2 as an example, the first number is 3, the object 1 is taken as a starting point, two edges connected to the object 1 exist, the probability of the two edges being selected is 50%, if an edge connected to the object 2 is selected, the object 2 is a second node in the supply and demand path, and so on, three edges connected to the object 2 exist, the probability of each edge being selected is 33.3%, if an edge connected to the object 3 is selected, the object 2 is a third node in the supply and demand path, and thus, the object 1, the object 2, and the object 3 are one supply and demand path. It should be noted that, in a supply and demand path, the same node may appear repeatedly, for example, the supply and demand path may also be object 1-object 2-object 1.
In the embodiment of the invention, the probability of selecting each edge connected with the same node is the same, so that the supply and demand path can uniformly cover each node, and the word vector model can uniformly learn the supply and demand characteristics of each object.
The embodiment of the invention determines the supply and demand paths by taking different nodes in the supply and demand relationship graph as the starting points respectively, so that the obtained supply and demand paths can cover all the nodes, and the accuracy of determining the target object is improved. Of course, in an actual application scenario, the supply and demand path may be determined only with a part of nodes as starting points.
Mode 2:
the supply and demand behavior data comprises: a transaction amount;
generating a plurality of supply and demand paths according to the supply and demand relationship diagram, wherein the method comprises the following steps:
for each node in the supply-demand relationship graph:
selecting edges connected with the nodes according to the weights of the edges by taking the nodes as starting points, and determining the nodes in the supply and demand paths according to the selected edges to obtain the supply and demand paths which are routed to the second number of nodes; wherein the weight of the edge is determined by the transaction amount between the nodes.
And the second number is the number of nodes included in the preset supply and demand path, and the first number is different from the second number in that the first number is applied to different scenes, the probability of each edge being selected is the same in the former scene, the weight of each edge exists in the latter scene, and the nodes in the supply and demand path are determined based on the weight of the edge.
In an embodiment of the invention, each edge has a corresponding weight determined by the transaction amount between the nodes. The transaction amount may be any one or more of the medium bid amount, the contract amount, and the invoice amount. Taking fig. 2 as an example, the second number is 3, the object 1 is taken as a starting point of the supply and demand path, two edges connected to the object 1 exist, the transaction amount of the object 1 and the object 2 is 25 ten thousand, the transaction amount of the object 1 and the object 4 is 75 ten thousand, and according to the proportion of the transaction amounts, the weight of the edge between the object 1 and the object 2 is determined to be 25%, the weight of the edge between the object 1 and the object 4 is determined to be 75%, and the object 4 is a second node of the supply and demand path because the weight of the edge between the object 1 and the object 4 is larger. If the transaction amount of the object 4 and the object 2 is 150 ten thousand, the weight of the edge between the object 4 and the object 2 is 66.7%, the weight of the edge between the object 4 and the object 1 is 33.3%, and the weight of the edge between the object 4 and the object 2 is larger, so the object 2 is the third node of the supply and demand path. Thus, the supply-demand path is object 1-object 4-object 2. It should be noted that the transaction amount between the nodes may be the accumulation of the transaction amounts between the nodes for a plurality of times, may also be the single transaction amount, and may also be the maximum transaction amount in a plurality of times of transactions. The supply and demand path may be formed by the name of the object or information such as a uniform social credit code of the object.
In the embodiment of the invention, the weight of the edge is determined through the transaction amount, different edges are distinguished, the probability of the node with higher transaction amount appearing in the supply and demand path is increased, the feature learning of the nodes is enhanced by the word vector model, and the accuracy of the similarity calculation result corresponding to the nodes is further ensured. In a practical application scenario, the weight of the edge may be determined according to one or more of the transaction amount, the transaction number, and the transaction time. Wherein, the larger the transaction frequency, the higher the weight of the edge, and the closer the transaction time, the higher the weight.
In one embodiment of the invention, generating an embedded representation of a plurality of objects according to a plurality of supply and demand paths comprises:
and inputting the multiple supply and demand paths into the word vector model to obtain the embedded representation of the multiple objects.
The embodiment of the invention inputs a word vector model through a plurality of supply and demand paths to train the word vector model and output the vector representation of each object. The embodiment of the invention vectorizes the supply and demand paths through the word vector model to obtain the embedded representation of each object, and can more quickly determine the target object based on the operation between the embedded representations. The Word vector model may be Word2vec, Global Vectors for Word reconstruction, etc.
And inputting each generated supply and demand path into word2vec to obtain the embedded representation of each object. For example, one of the supply and demand paths is "Beijing AA technologies, Inc. -Beijing BB network technologies, Inc.".
The embedded representation of the object can be a vector of arbitrary dimensions. Taking a 16-dimensional vector as an example:
the "Beijing AA technologies Co., Ltd" insert is expressed as: [ -0.6433484, 1.9626732, 2.9946766, 3.4748187, 0.8176478, -0.945684, 1.0036267, 1.8913803, 1.430759, 1.2809728, 4.0172596, 2.8226984, -1.9158391, 0.17588441, -3.302099, 1.3402888].
The embedded expression of "Beijing BB network technology Co., Ltd" is: [ -0.52615666, -1.8757683, 2.3022957, -2.247738, 3.6796074, 0.26537383, 1.8951517, -0.5244883, 0.4057679, 3.4313507, -0.7072354, -4.1955266, -1.4017067, 1.5180964, -3.0574412, 1.4780037].
The similarity of the embedded representations of the two objects characterizes the co-occurrence of the two objects in the supply and demand paths, and when the two objects appear in different supply and demand paths for multiple times, the similarity of the embedded representations of the two objects is higher than the similarity of the embedded representations of the two objects which do not appear simultaneously.
In the practical application scenario, at least two ways of determining the target object are included.
Mode 1:
in one embodiment of the present invention, determining a target object having a similar supply chain as the current object based on the embedded representation of the respective object comprises:
determining a similarity of the embedded representation of the current object to the embedded representations of the respective other objects;
selecting a third number of objects from the plurality of other objects as target objects according to the sequence that the similarity of the embedded representation of the current object is from large to small; and the similarity between the embedded representation of the target object and the embedded representation of the current object is used for representing the similarity between the supply chain of the target object and the supply chain of the current object.
In the embodiment of the present invention, the similarity between the embedded representation of the target object and the embedded representation of the current object is the similarity between the supply chain of the target object and the supply chain of the current object. The similarity between the embedded representation of the current object and the embedded representations of other objects can be calculated by the distance between the two, such as cosine distance, Euclidean distance, etc., and the larger the distance, the smaller the similarity. For example, the similarity of the embedded representation of the current object to the embedded representations of other objects is the inverse of the Euclidean distance between them, or "1-cosine distance". It is an aim of embodiments of the present invention to select a third number of target objects from a plurality of other objects that are most similar to the supply chain of the current object. The embodiment of the invention converts the similarity of the determined supply chain into the similarity of the determined embedded expression, namely, the similarity is converted into the operation among vectors, so that the final target object can be obtained more quickly.
Mode 2:
determining target objects having a similar supply chain as the current object based on the embedded representations of the respective objects, comprising:
adding the current object to the search set and the recall set;
according to the first-in first-out sequence, first objects ranked at the first position are obtained from a search set, and a plurality of second objects with similarity larger than a first threshold value with the first objects are searched;
for each second object: terminating the current flow if the second object exists in the recall set; adding the second object to the search set and the recall set if the second object does not exist in the recall set;
removing the first object from the corpus;
determining whether the number of the second objects in the recall set is greater than the fourth number or whether the search set is empty, and if so, determining that the second objects in the recall set are target objects; otherwise, acquiring the first objects ranked at the first position from the search set according to the first-in first-out sequence, and searching a plurality of second objects with the similarity to the first objects being greater than a first threshold value;
and the similarity between the embedded representation of the target object and the embedded representation of the current object is used for representing the similarity between the supply chain of the target object and the supply chain of the current object. A first threshold for determining the range of each search, the value of which is set by the user in dependence on the distance between the embedded representations of the objects, the first threshold being increased accordingly if the distance between the embedded representations of the objects is larger, and the first threshold being decreased accordingly if the distance between the embedded representations of the objects is smaller. For example, the number of target objects to be determined is 100, the first threshold is 0.98, at this time, the number of second objects obtained by the first search is 10, the first threshold is reduced to 0.95, the number of second objects obtained by the first search is 30, and the user may determine the specific value of the first threshold according to the requirement. The fourth number is used to define a range of numbers of target objects, and in the embodiment of the present invention, the number of target objects to be acquired is larger than the fourth number.
For example, the first threshold value is set to 90% and the fourth number is set to 7 in advance. There are 9 objects, A, B, C, D, E, F, G, H and I, a being current objects, and a is added to the search and recall sets, where there is only a in both the search and recall sets, the first ranked a is obtained from the search set, and objects with similarity greater than 90% (first threshold) to the embedded representation of a are searched among the other objects (B, C, D, E, F, G, H and I) except a, resulting in C, D, E, F and G. And judging C, determining whether C exists in the recall set, and adding C to the search set and the recall set because C is not in the recall set. By analogy, D, E, F and G were judged in turn, and since both D, E, F and G were absent from the recall, D, E, F and G were added to the search set and recall. Remove a from the corpus. At this time, the search set includes: C. d, E, F and G. The recall set includes: A. c, D, E, F and G, in amounts less than 7. C ranked first is acquired from the search set in a first-in-first-out order, and objects having a similarity greater than 90% (first threshold) to the embedded representation of C are searched among the objects (A, B, D, E, F, G, H and I) other than C. And repeating the steps until the number of the objects in the recall set is greater than 7 or the search set is empty, and determining the objects in the recall set as target objects.
According to the embodiment of the invention, the objects in the same cluster can be preferentially obtained by searching layer by layer through the set first threshold value, so that the obtained target objects are more fit with the actual service scene. Because the transaction is easier to generate between objects located in the same region compared with objects located in different regions, the embedded representation of the objects may be clustered, for example, enterprises located in the province of Hunan province are clustered into a cluster 1, and enterprises located in the province of Hubei province are clustered into a cluster 2, if the current enterprise is located in the province of Hunan province, the enterprise in the province of Hunan province is preferentially obtained through the method 2 provided by the embodiment of the invention, and the result more meets the requirements of actual business scenarios.
As shown in fig. 3, an embodiment of the present invention provides a method for determining enterprises with similar supply chains, including:
step 301: bid data, contract data, and invoice data for a plurality of enterprises are obtained.
Step 302: and determining the enterprises with supply and demand relations according to the bidding data, contract data and invoice data of each enterprise.
Step 303: and generating a supply and demand relation graph by taking the enterprises with the supply and demand relation as nodes and the supply and demand relation as edges.
Step 304: for each node in the supply-demand relationship graph: and randomly selecting edges connected with the nodes by taking the nodes as a starting point, and determining the nodes in the supply and demand paths according to the selected edges to obtain the supply and demand paths passing through the first number of nodes.
Wherein, the supply and demand path comprises a plurality of enterprises.
Step 305: and inputting a plurality of supply and demand paths into word2vec to obtain embedded representations of a plurality of enterprises.
Step 306: a euclidean distance of the embedded representation of the current business from the embedded representations of the respective other businesses in the plurality of businesses is determined.
The similarity between the embedded representation of the current enterprise and the embedded representations of other enterprises can be characterized, and therefore, the Euclidean distance between the embedded representation of the current enterprise and the embedded representations of other enterprises can be characterized.
Step 307: a third number of businesses are selected from the plurality of other businesses as target businesses in descending order of euclidean distance from the embedded representation of the current business.
And the similarity between the supply chain of the target enterprise and the supply chain of the current enterprise is not less than a first threshold value.
The embodiment of the invention is based on the bidding data, contract data and invoice data of enterprises, considers the supply and demand relations of different types among the enterprises, constructs a more comprehensive supply and demand relation graph, and extracts the supply and demand paths from the supply and demand relation graph. In addition, the method converts the supply and demand paths into the embedded representation, and can determine the target enterprise more quickly through vector operation.
As shown in fig. 4, an embodiment of the present invention provides an apparatus for determining a target object, including:
an embedded representation determining module 401 configured to determine embedded representations of a plurality of objects according to a predetermined supply-demand relationship diagram;
an object determination module 402 configured to determine target objects having a similar supply chain as the current object based on the embedded representations of the respective objects; wherein the plurality of objects includes: the current object and the target object; the supply and demand relation graph takes the objects with the supply and demand relation as nodes and the supply and demand relation as edges, and the objects with the supply and demand relation are determined by the supply and demand behavior data of the objects.
In one embodiment of the invention, the apparatus further comprises: a graph determination module;
the graph determining module is configured to obtain the supply and demand behavior data of each object; determining objects with supply and demand relations according to the supply and demand behavior data of each object; and generating a supply and demand relation graph by taking the object with the supply and demand relation as a node and the supply and demand relation as an edge.
In an embodiment of the present invention, the embedded representation determining module 401 is configured to generate a plurality of supply and demand paths according to the supply and demand relationship diagram; generating embedded representations of a plurality of objects according to a plurality of supply and demand paths; wherein, the supply and demand path comprises a plurality of objects.
In one embodiment of the invention, the embedded representation determination module 401 is configured to, for each node in the supply-demand relationship graph: and randomly selecting edges connected with the nodes by taking the nodes as a starting point, and determining the nodes in the supply and demand paths according to the selected edges to obtain the supply and demand paths passing through the first number of nodes.
In one embodiment of the invention, the supply and demand behavior data comprises: a transaction amount;
an embedded representation determination module 401 configured to, for each node in the supply-demand relationship graph: selecting edges connected with the nodes according to the weights of the edges by taking the nodes as starting points, and determining the nodes in the supply and demand paths according to the selected edges to obtain the supply and demand paths which are routed to the second number of nodes; wherein the weight of the edge is determined by the transaction amount between the nodes.
In one embodiment of the invention, the embedded representation determining module 401 is configured to input a plurality of supply and demand paths into the word vector model, resulting in an embedded representation of a plurality of objects.
In one embodiment of the invention, the object determination module 402 is configured to determine a similarity of the embedded representation of the current object to the embedded representations of the respective other objects; selecting a third number of objects from the plurality of other objects as target objects according to the sequence of the similarity of the embedded representation of the current object from large to small; and the similarity between the embedded representation of the target object and the embedded representation of the current object is used for representing the similarity between the supply chain of the target object and the supply chain of the current object.
In one embodiment of the invention, the object determination module 402 is configured to add the current object to the search and recall set; according to the first-in first-out sequence, first objects ranked at the first position are obtained from a search set, and a plurality of second objects with the similarity of the embedded representation of the first objects larger than a first threshold value are searched in the plurality of objects; for each second object: terminating the current flow if the second object exists in the recall set; adding the second object to the search set and the recall set if the second object does not exist in the recall set; removing the first object from the corpus; determining whether the number of the second objects in the recall set is greater than the fourth number or whether the search set is empty, and if so, determining that the second objects in the recall set are target objects; otherwise, acquiring a first object arranged at the first position from the search set according to the first-in first-out sequence, and searching a plurality of second objects with the similarity of the embedded representation of the first object being greater than a first threshold value; and the similarity between the embedded representation of the target object and the embedded representation of the current object is used for representing the similarity between the supply chain of the target object and the supply chain of the current object.
With regard to the apparatus in the above-described embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.
Based on the same inventive concept, embodiments of the present invention also provide a non-transitory computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, implements the steps of the above method for determining a target object.
Specifically, the computer-readable storage medium may be a flash memory, a hard disk, a multimedia card, a card type memory (e.g., SD or DX memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a Read Only Memory (ROM), an Electrically Erasable Programmable Read Only Memory (EEPROM), a Programmable Read Only Memory (PROM), a magnetic memory, a magnetic disk, an optical disk, a server, a public cloud server, etc.
With regard to the computer-readable storage medium in the above-described embodiments, the method steps for determining the target object when the computer program stored thereon is executed will be described in detail in relation to the embodiments of the method, and will not be elaborated upon here.
Based on the same inventive concept, an embodiment of the present invention further provides an electronic device, including:
a memory having a computer program stored thereon;
a processor for executing a computer program in a memory for carrying out the steps of the method of determining a target object as described above.
Fig. 5 is a block diagram illustrating an electronic device 500 in accordance with an example embodiment. As shown in fig. 5, the electronic device 500 may include: a processor 501 and a memory 502. The electronic device 500 may also include one or more of a multimedia component 503, an input/output (I/O) interface 504, and a communication component 505.
The processor 501 is configured to control the overall operation of the electronic device 500, so as to complete all or part of the steps in the method for determining a target object. The memory 502 is used to store various types of data to support operation at the electronic device 500, such as instructions for any application or method operating on the electronic device 500 and application-related data, such as contact data, messaging, pictures, audio, video, and so forth. The Memory 502 may be implemented by any type of volatile or non-volatile Memory device or combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic Memory, flash Memory, magnetic disk or optical disk. The multimedia component 503 may include a screen and an audio component. Wherein the screen may be, for example, a touch screen and the audio component is used for outputting and/or inputting audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may further be stored in the memory 502 or transmitted through the communication component 505. The audio assembly further comprises at least one speaker for outputting audio signals. The I/O interface 504 provides an interface between the processor 501 and other interface modules, such as a keyboard, mouse, buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 505 is used for wired or wireless communication between the electronic device 500 and other devices. Wireless Communication, such as Wi-Fi, bluetooth, Near Field Communication (NFC), 2G, 3G, 4G or 5G, NB-IOT (Narrow Band Internet of Things), or a combination of one or more of them, so that the corresponding Communication component 505 may include: Wi-Fi module, bluetooth module, NFC module.
In an exemplary embodiment, the electronic Device 500 may be implemented by one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components for performing the above-described method for determining a target object.
The preferred embodiments of the present invention have been described in detail with reference to the accompanying drawings, however, the present invention is not limited to the specific details of the above embodiments, and various simple modifications can be made to the technical solution of the present invention within the technical idea of the present invention, and these simple modifications are within the protective scope of the present invention.
It should be noted that the various technical features described in the above embodiments can be combined in any suitable manner without contradiction, and the invention is not described in any way for the possible combinations in order to avoid unnecessary repetition.
In addition, any combination of the various embodiments of the present invention is also possible, and the same should be considered as the disclosure of the present invention as long as it does not depart from the spirit of the present invention.
Claims (7)
1. A method of determining a target object, comprising:
determining word vector representations of a plurality of objects according to a predetermined supply-demand relationship diagram, comprising:
generating a plurality of supply and demand paths according to the supply and demand relationship diagram; wherein the supply and demand path comprises a plurality of objects; inputting the multiple supply and demand paths into a word vector model to obtain word vector representations of the multiple objects;
determining a target object having a similar supply chain as the current object based on the word vector representation of each of the objects, comprising:
determining similarity of the word vector representation of the current object to the word vector representations of the respective other objects;
selecting a third number of objects from a plurality of other objects as the target objects according to the sequence that the similarity represented by the word vector of the current object is from large to small; wherein the word vector of the target object represents the similarity to the word vector representation of the current object for characterizing the similarity of the supply chain of the target object to the supply chain of the current object;
wherein the plurality of objects includes: the current object and the target object; the supply and demand relation graph takes an object with a supply and demand relation as a node and the supply and demand relation as an edge, and the object with the supply and demand relation is determined by the supply and demand behavior data of the objects.
2. The method of claim 1,
generating a plurality of supply and demand paths according to the supply and demand relationship diagram, wherein the generating comprises:
for each node in the supply-demand relationship graph:
and randomly selecting edges connected with the nodes by taking the nodes as a starting point, and determining the nodes in the supply and demand paths according to the selected edges to obtain the supply and demand paths passing through the nodes of the first number.
3. The method of claim 1,
the supply and demand behavior data comprises: a transaction amount;
generating a plurality of supply and demand paths according to the supply and demand relationship diagram, wherein the generating comprises:
for each node in the supply-demand relationship graph:
selecting edges connected with the nodes according to the weights of the edges by taking the nodes as starting points, and determining the nodes in the supply and demand paths according to the selected edges to obtain the supply and demand paths which are routed to the second number of nodes; wherein the weight of the edge is determined by the transaction amount between the nodes.
4. The method of claim 1,
said determining a target object having a similar supply chain as the current object based on the word vector representation of each of said objects comprises:
adding the current object to a search set and recall set;
according to the first-in first-out sequence, first objects ranked at the first position are obtained from the search set, and a plurality of second objects with similarity larger than a first threshold value and represented by word vectors of the first objects are searched in the plurality of objects;
for each of the second objects: terminating the current flow if the second object is present in the recall set; adding the second object to the search set and the recall set if the second object is not present in the recall set;
deleting the first object from the corpus;
determining whether the number of second objects in the recall set is greater than a fourth number or whether the search set is empty, and if so, determining that the second objects in the recall set are the target objects; otherwise, executing the first-in first-out sequence, acquiring a first object ranked at the first position from the search set, and searching a plurality of second objects of which the similarity represented by the word vector of the first object is greater than a first threshold value;
and the word vector of the target object represents the similarity of the word vector of the current object, and is used for representing the similarity of the supply chain of the target object and the supply chain of the current object.
5. An apparatus for determining a target object, comprising:
the word vector representation determining module is configured to generate a plurality of supply and demand paths according to the supply and demand relation diagram; inputting the multiple supply and demand paths into a word vector model to obtain word vector representations of multiple objects; wherein, the supply and demand path comprises a plurality of objects;
an object determination module configured to determine a similarity of the word vector representation of the current object to the word vector representations of the respective other objects; selecting a third number of objects from the plurality of other objects as target objects according to the sequence that the similarity represented by the word vector of the current object is from large to small; the word vector of the target object represents the similarity of the word vector of the current object, and is used for representing the similarity of the supply chain of the target object and the supply chain of the current object;
wherein the plurality of objects includes: the current object and the target object; the supply and demand relation graph takes an object with a supply and demand relation as a node and the supply and demand relation as an edge, and the object with the supply and demand relation is determined by the supply and demand behavior data of the objects.
6. A non-transitory computer-readable storage medium, on which a computer program is stored, which program, when executed by a processor, performs the steps of the method of any one of claims 1 to 4.
7. An electronic device, comprising:
a memory having a computer program stored thereon;
a processor for executing the computer program in the memory to carry out the steps of the method of any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210117160.9A CN114155049B (en) | 2022-02-08 | 2022-02-08 | Method and device for determining target object |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210117160.9A CN114155049B (en) | 2022-02-08 | 2022-02-08 | Method and device for determining target object |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114155049A CN114155049A (en) | 2022-03-08 |
| CN114155049B true CN114155049B (en) | 2022-05-17 |
Family
ID=80450253
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210117160.9A Active CN114155049B (en) | 2022-02-08 | 2022-02-08 | Method and device for determining target object |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114155049B (en) |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9779135B2 (en) * | 2011-11-03 | 2017-10-03 | Sap Se | Semantic related objects |
| CN110807103A (en) * | 2019-10-18 | 2020-02-18 | 中国银联股份有限公司 | Knowledge graph construction method and device, electronic equipment and storage medium |
| CN111476668B (en) * | 2020-06-24 | 2020-10-13 | 支付宝(杭州)信息技术有限公司 | Identification method and device of credible relationship, storage medium and computer equipment |
| CN113222771B (en) * | 2020-07-10 | 2023-10-20 | 杭州海康威视数字技术股份有限公司 | Method and device for determining target group based on knowledge graph and electronic equipment |
| CN112989169B (en) * | 2021-02-23 | 2023-07-25 | 腾讯科技（深圳）有限公司 | Target object identification method, information recommendation method, device, equipment and medium |
-
2022
- 2022-02-08 CN CN202210117160.9A patent/CN114155049B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN114155049A (en) | 2022-03-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20180336193A1 (en) | Artificial Intelligence Based Method and Apparatus for Generating Article | |
| WO2017193897A1 (en) | Data recommendation method and device therefor, and storage medium | |
| US20190340688A1 (en) | Utilizing artificial intelligence to make a prediction about an entity based on user sentiment and transaction history | |
| WO2019114423A1 (en) | Method and apparatus for merging model prediction values, and device | |
| US10657543B2 (en) | Targeted e-commerce business strategies based on affiliation networks derived from predictive cognitive traits | |
| CN109471978B (en) | Electronic resource recommendation method and device | |
| CN108153909B (en) | Keyword putting word-expanding method and device, electronic equipment and storage medium | |
| CN110782286A (en) | Advertisement pushing method and device, server and computer readable storage medium | |
| US20190102784A1 (en) | Modeling sequential actions | |
| CN113722546B (en) | Abnormal user account acquisition method and device, electronic equipment and storage medium | |
| CN114978484A (en) | Data processing method and device for protecting privacy and computer equipment | |
| US20200098031A1 (en) | Product recommending apparatus and non-transitory computer readable medium | |
| CN111782925A (en) | Item recommendation method, device, equipment, system and readable storage medium | |
| CN114119123A (en) | Information pushing method and device | |
| CN116823410B (en) | Data processing method, object processing method, recommending method and computing device | |
| US10579752B2 (en) | Generating a model based on input | |
| CN112650942A (en) | Product recommendation method, device, computer system and computer-readable storage medium | |
| US20230325857A1 (en) | Method and system of sentiment-based selective user engagement | |
| CN114155049B (en) | Method and device for determining target object | |
| CN113744002A (en) | Method, device, equipment and computer readable medium for pushing information | |
| CN111787042A (en) | Method and device for pushing information | |
| CN110971973A (en) | Video pushing method and device and electronic equipment | |
| CN110609783A (en) | Method and device for identifying abnormal behavior user | |
| Stiebellehner et al. | Learning continuous user representations through hybrid filtering with doc2vec | |
| Zhang et al. | Service2vec: A vector representation for web services |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |