WO2023035940A1

WO2023035940A1 - Target object recommendation method and system

Info

Publication number: WO2023035940A1
Application number: PCT/CN2022/114376
Authority: WO
Inventors: 姚娟娟; 钟南山
Original assignee: 上海明品医学数据科技有限公司
Priority date: 2021-09-10
Filing date: 2022-08-24
Publication date: 2023-03-16
Also published as: CN113505304B; CN113505304A

Abstract

The present invention provides a target object recommendation method and system. The target object recommendation method comprises: obtaining target texts of target objects to be assigned, obtaining multiple feature tags by means of the target texts, and obtaining feature matrices by means of the multiple feature tags; inputting the feature matrices into a neural network for classification processing to obtain a classification result, and obtaining a recommendation model by training the neural network, the classification result comprising classified contracting objects and first confidence levels; and inputting the feature matrices corresponding to multiple target objects to be assigned into the recommendation model, respectively obtaining recommended classification results, and determining, according to the recommended classification results, the target objects to be assigned to the contracting objects.

Description

A target object recommendation method and system

technical field

The present invention relates to the technical field of big data, in particular to a target object recommendation method and system.

Background technique

With the development of economy and society, the general public's demand for services is becoming more and more specialized and subdivided. Limited by the ways and means of obtaining services, service providers often cannot find accurate and ideal target objects, especially in Among the highly professional contracting services, the contracting service refers to the behavior that the party/contractor completes a certain work for the target party, and the target party pays the agreed remuneration after acceptance.

At present, the purpose of allocating target objects can be achieved by building an informationized service platform to dispatch orders for contracting objects. However, when the target object sends a request on the service platform, how can the platform accurately dispatch the appropriate target object to the warehouse? Dealing with the contracted object will be a practical problem, which directly affects the sustainable development of the service platform, for example, the satisfaction degree of the direct target object and the stickiness of the contracted object.

Contents of the invention

In view of the above-mentioned shortcomings of the prior art, the purpose of the present invention is to provide a target object recommendation method and system for solving the problem of inaccurate target object recommendation in the prior art.

In order to achieve the above purpose and other related purposes, an embodiment of the present invention provides a method for recommending a target object, including:

Obtain the target text of the target object to be assigned, obtain a plurality of feature tags through the target text, and obtain a feature matrix through a plurality of the feature tags;

Inputting the feature matrix into the neural network for classification processing, obtaining classification results, and obtaining a recommendation model by training the neural network, the classification results including classified contract objects and first confidence;

Inputting feature matrices corresponding to multiple target objects to be assigned into the recommendation model, obtaining recommended classification results respectively, and determining target objects to be assigned by the contracting object according to the recommended classification results.

Further, determining the target object to be assigned to the contract object according to the recommended classification result includes:

A first threshold is set, and when the first confidence level is greater than or equal to the first threshold, it is determined that the recommendation is valid, and the target object to be assigned is assigned to the contract object.

Further, a plurality of feature tags are obtained through the target text, including:

Performing natural language processing on the text of the target object to be assigned, comparing the processing result with the marked feature labels, and obtaining the processed feature labels according to the similarity.

Further, the neural network includes an input layer, a hidden layer and an output layer, and the neural network is trained through a loss function, the loss function includes a first loss function and a second loss function, and the first loss function includes The loss between the processed feature label and the marked feature label, the second loss function includes the loss between the target object to be assigned and the recommended contract object.

Further, the mathematical expression of the loss function is:

L=L1+L2

Wherein, L is a loss function, L1 is a first loss function, L2 is a second loss function, N is the number of processed feature labels, and M is the number of contract labels;

When the i-th label in the set of processed feature labels matches the e-th label in the set of marked feature labels, x(ie)=1, otherwise, x(ie)=0;

Pie is the probability that the i-th label in the set of processed feature labels matches the e-th label in the set of marked feature labels;

When the kth target object to be assigned matches the recommended contract object, zk=1, otherwise, zk=0;

pk is the probability of the target object to be assigned and the recommended contract object.

Further, after determining the target object to be allocated by the contracting object according to the recommended classification result, it includes:

When the contracting object rejects the target object to be allocated, changing the first confidence level of the contracting object to obtain a second confidence level;

When the second degree of confidence is greater than or equal to the first threshold, it is determined that the recommendation is valid, and the target object to be assigned is assigned to the contract object;

When the second confidence level is smaller than the first threshold, the target object to be allocated is allocated to the contracting object with the highest first confidence level.

Further, the mathematical expression of the second confidence degree is:

Wherein, t is the number of rejections of the contracted objects per unit time, b is the average number of rejections of each of the contracted objects per unit time, e is the natural logarithm, a1 is the first degree of confidence, and a2 is the second degree of confidence .

A target object recommendation system, comprising:

The acquisition module acquires the target text of the target object to be assigned, obtains a plurality of feature tags through the target text, and obtains a feature matrix through a plurality of the feature tags;

A model module, configured to input the feature matrix into the neural network for classification processing, obtain classification results, and obtain a recommendation model by training the neural network, the classification results include classified contract objects and first confidence;

A processing module, configured to input feature matrices corresponding to a plurality of target objects to be allocated into the recommendation model, obtain recommended classification results respectively, and determine that the contracting object will The assigned target object.

An electronic device comprising:

one or more processors; and

One or more machine-readable media having instructions stored thereon, when executed by the one or more processors, causes the electronic device to execute the target object recommendation method.

A machine-readable medium, on which instructions are stored, when executed by one or more processors, causes the device to execute the target object recommendation method.

Description of drawings

FIG. 1 is a schematic diagram of a target object recommendation method according to one embodiment of the present invention.

FIG. 2 is a schematic diagram of a target object recommendation system according to one embodiment of the present invention.

Detailed ways

Embodiments of the present invention are described below through specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific implementation modes, and various modifications or changes can be made to the details in this specification based on different viewpoints and applications without departing from the spirit of the present invention.

It should be noted that the diagrams provided in this embodiment are only schematically illustrating the basic idea of the present invention, and only the components related to the present invention are shown in the diagrams rather than the number, shape and shape of the components in actual implementation. Dimensional drawing, the type, quantity and proportion of each component can be changed arbitrarily during actual implementation, and the component layout type may also be more complicated. The structures, proportions, sizes, etc. shown in the drawings attached to this specification are only used to match the content disclosed in the specification for the understanding and reading of those who are familiar with this technology, and are not used to limit the conditions for the implementation of the present invention , so it has no technical substantive meaning, and any modification of structure, change of proportional relationship or adjustment of size shall still fall within the scope of the disclosure of the present invention without affecting the functions and objectives of the present invention. The technical content must be within the scope covered. At the same time, terms such as "upper", "lower", "left", "right", "middle" and "one" quoted in this specification are only for the convenience of description and are not used to limit this specification. The practicable scope of the invention and the change or adjustment of its relative relationship shall also be regarded as the practicable scope of the present invention without any substantial change in the technical content.

In the contracting service recommendation platform, the target object sends a demand request to the platform, and the platform recommends the appropriate target object to the specific contracting object according to the characteristics of the contracting object, which not only can meet the needs of the target object, but also the contracting object can obtain the ideal target User, reduce opportunity cost, please refer to Figure 1, the present invention provides a method for recommending target objects, including:

S1: Obtain the target text of the target object to be assigned, obtain a plurality of feature tags through the target text, and obtain a feature matrix through the plurality of feature tags;

S2: Input the feature matrix into the neural network for classification processing, obtain the classification result, and obtain the recommendation model by training the neural network, the classification result includes the classified contract object and the first confidence degree, in the classification processing During the process, the characteristic matrix can be subjected to matrix operation to obtain the corresponding relationship between the characteristic matrix and the contractor;

S3: Input feature matrices corresponding to multiple target objects to be assigned into the recommendation model, obtain recommended classification results respectively, and determine the target to be assigned by the contracting object according to the recommended classification results Objects, through the classification process of the recommendation model, obtain the classified contract objects and the first confidence degree, select the classified contract objects with higher first confidence degree according to the size of the first confidence degree, and determine the target to be assigned The objects are matched with the classified contracting objects with a higher first degree of confidence and allocated, which improves the accuracy of allocation of potential target objects and improves the satisfaction of contracting services.

In some implementations, multiple feature tags are obtained through the target text, including:

Setting the first threshold, when the first confidence is greater than or equal to the first threshold, it is determined that the recommendation is valid, and the target object to be assigned is assigned to the contract object, which improves the efficiency and accuracy of the recommendation and avoids The assigned target object is assigned to the contract object with a lower first confidence level.

In order to obtain the feature label, a natural language processing method can be used, for example, to obtain the feature label of the target object to be assigned, including:

In some implementations, the neural network includes an input layer, a hidden layer, and an output layer, and the neural network is trained through a loss function, and the loss function includes a first loss function and a second loss function, and the first The loss function includes a loss between the processed feature label and the marked feature label, and the second loss function includes a loss between the target object to be assigned and the recommended contract object.

Further, the mathematical expression of the loss function is:

L=L1+L2

pk is the probability of the target object to be assigned and the recommended contract object. The loss function not only considers the loss between the processed feature label and the marked feature label, but also considers the loss between the target object to be assigned and the recommended contract object, which improves the classification process. The recommended accuracy in .

In order to improve the processing efficiency of the target objects to be allocated and avoid frequent rejection of the target objects to be allocated by the contracting object, the ability to change the opportunity to obtain the target objects to be allocated can be set up according to the number or frequency of rejection recommendations, for example, according to the After the classification results after the above recommendations determine the target objects to be assigned by the contract objects, it includes:

When the second confidence degree is greater than or equal to the first threshold, it is determined that the recommendation is valid, and the target object to be assigned is assigned to the contract object;

Further, the mathematical expression of the second confidence degree is:

Wherein, t is the number of rejections of the contracted objects per unit time, b is the average number of rejections of each of the contracted objects per unit time, e is the natural logarithm, a1 is the first degree of confidence, and a2 is the second degree of confidence . For example, when t=1 and b=2, it means that the frequency of rejection by the contracted object is lower than the average number, so it can properly improve its ability to obtain the target object to be assigned, then a2>a1; when t=3, b= When 2, it means that the frequency of rejection by the contracted object is greater than the average number, so the opportunity ability to obtain the target object to be assigned can be appropriately reduced, then a2<a1.

Referring to Fig. 2, one of the embodiments of the present invention also provides a target object recommendation system, including:

An acquisition module that acquires the target text of the target object to be assigned, acquires a plurality of feature tags through the target text, and acquires a feature matrix through a plurality of the feature tags;

Further, determining the target object to be assigned to the contracted object according to the recommended classification result includes:

Further, the feature label of the target object to be assigned is obtained, including:

Further, the mathematical expression of the loss function is:

L=L1+L2

Further, after determining the target object to be allocated by the contract object according to the recommended classification result, it includes:

Further, the mathematical expression of the second confidence degree is:

An embodiment of the present invention provides an electronic device, including: one or more processors; and one or more machine-readable media storing instructions thereon, and when the one or more processors are executed, the An electronic device performs one or more of the described methods. The invention is applicable to numerous general purpose and special purpose computing system environments or configurations. Examples: personal computers, server computers, handheld or portable devices, tablet-type devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, including A distributed computing environment for any of the above systems or devices, etc.

The embodiments of the present invention also provide one or more machine-readable media, on which are stored instructions, which, when executed by one or more processors, cause the device to perform one or more of the above-mentioned methods. The invention may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including storage devices.

As mentioned above, the target object recommendation method and system of the present invention have the following beneficial effects:

Through the classification process of the recommendation model, the classified contract objects and the first confidence degree are obtained, and according to the size of the first confidence degree, the classified contract objects with a higher first confidence degree are selected, and the target object to be assigned is determined to be related to The contracted objects classified after the first confidence level are matched and allocated, which improves the accuracy of allocation of potential target objects and improves the satisfaction of contracted services.

The above-mentioned embodiments only illustrate the principles and effects of the present invention, but are not intended to limit the present invention. Anyone skilled in the art can modify or change the above-mentioned embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or changes made by those skilled in the art without departing from the spirit and technical ideas disclosed in the present invention should still be covered by the claims of the present invention.

Claims

A target object recommendation method, characterized in that it includes:

Obtain the target text of the target object to be assigned, obtain a plurality of feature tags through the target text, and obtain a feature matrix through a plurality of the feature tags;

Inputting the feature matrix into the neural network for classification processing, obtaining classification results, and obtaining a recommendation model by training the neural network, the classification results including classified contract objects and first confidence;

Inputting feature matrices corresponding to multiple target objects to be assigned into the recommendation model, obtaining recommended classification results respectively, and determining target objects to be assigned by the contracting object according to the recommended classification results.
The target object recommendation method according to claim 1, wherein determining the target object to be assigned to the contract object according to the recommended classification result includes:

A first threshold is set, and when the first confidence level is greater than or equal to the first threshold, it is determined that the recommendation is valid, and the target object to be assigned is assigned to the contract object.
The target object recommendation method according to claim 2, wherein obtaining a plurality of feature labels through the target text includes:

Performing natural language processing on the text of the target object to be assigned, comparing the processing result with the marked feature labels, and obtaining the processed feature labels according to the similarity.
The target object recommendation method according to claim 3, wherein the neural network includes an input layer, a hidden layer, and an output layer, and the neural network is trained through a loss function, and the loss function includes a first loss function and a second loss function, the first loss function includes the loss between the processed feature label and the labeled feature label, and the second loss function includes the target object to be assigned and the recommended contract object loss between.
The target object recommendation method according to claim 4, wherein the mathematical expression of the loss function is: L=L1+L2

Wherein, L is a loss function, L1 is a first loss function, L2 is a second loss function, N is the number of processed feature labels, and M is the number of contract labels;

When the i-th label in the set of processed feature labels matches the e-th label in the set of marked feature labels, x(ie)=1, otherwise, x(ie)=0;

Pie is the probability that the i-th label in the set of processed feature labels matches the e-th label in the set of marked feature labels;

When the kth target object to be assigned matches the recommended contract object, zk=1, otherwise, zk=0;

pk is the probability of the target object to be assigned and the recommended contract object.
The target object recommendation method according to claim 1, characterized in that, after determining the target object to be allocated by the contract object according to the recommended classification result, comprising:

When the contracting object rejects the target object to be allocated, changing the first confidence level of the contracting object to obtain a second confidence level;

When the second confidence degree is greater than or equal to the first threshold, it is determined that the recommendation is valid, and the target object to be assigned is assigned to the contract object;

When the second confidence level is smaller than the first threshold, the target object to be allocated is allocated to the contracting object with the highest first confidence level.
The method for recommending target objects according to claim 6, wherein the mathematical expression of the second degree of confidence is:

Among them, t is the number of rejections of the contracted objects per unit time, b is the average number of rejections of each of the contracted objects per unit time, e is the natural logarithm, a1 is the first degree of confidence, and a2 is the second degree of confidence .
A target object recommendation system, characterized in that it comprises:

The acquisition module acquires the target text of the target object to be assigned, obtains a plurality of feature tags through the target text, and obtains a feature matrix through a plurality of the feature tags;

The model module is used to input the feature matrix into the neural network to perform classification processing, obtain the classification result, and obtain the recommendation model by training the neural network, and the classification result includes the classified contract object and the first degree of confidence;

A processing module, configured to input feature matrices corresponding to a plurality of target objects to be allocated into the recommendation model, obtain recommended classification results respectively, and determine that the contracting object will The assigned target object.
An electronic device, characterized in that it comprises:

One or more processors; and one or more machine-readable media having instructions stored thereon, which, when executed by the one or more processors, cause the electronic device to perform any one of claims 1-7 The target object recommendation method described above.
A machine-readable medium, characterized in that instructions are stored thereon, and when executed by one or more processors, the device executes the method for recommending a target object according to any one of claims 1-7.