CN107844980A - Commercial articles true and false discrimination method and device, computer-readable storage medium and equipment - Google Patents

Abstract

The invention relates to a commodity authenticity identification method and device, a computer storage medium and equipment, which obtain a commodity authenticity identification model determined by machine learning according to a feature information database of a preset sample commodity through a machine learning method, and obtain the authenticity of the commodity to be tested. Feature information, using the feature information of the product to be tested as the input of the product authenticity identification model to identify the authenticity of the product, that is, using the product authenticity identification model established by the preset machine learning method based on the feature information database of the sample product, and the product to be tested Carry out true and false identification, so as to realize the accurate identification of the true and false of the product to be tested, and do not need to manually identify the true and false, avoid the problem of affecting the identification result due to artificial subjectivity, improve the accuracy of product identification, and then according to the identification data and The data ranges corresponding to each preset authenticity level determine the authenticity level of the commodity to be tested, so as to know the authenticity degree of the commodity to be tested more intuitively.

Description

Commodity authenticity identification method and device, computer storage medium and equipment

technical field

The invention relates to the technical field of data processing, in particular to a method and device for authenticating a commodity, a computer storage medium and equipment.

Background technique

With the rapid development of e-commerce, online shopping has entered thousands of households, and people can enjoy the convenience of shopping in shopping malls without leaving home. Due to the virtuality of the network and the limitations of the relevant regulatory system, the goods on the e-commerce platform are uneven, filled with a large number of fake and shoddy products, which not only damages the rights and interests of consumers and regular manufacturers, but also brings choices to consumers when shopping online. The difficulty of commodities has seriously affected the healthy development of e-commerce. Therefore, in order to ensure the healthy development of e-commerce, it is very necessary to identify the authenticity of commodities.

At present, the common method for identifying the authenticity of commodities is for consumers to manually identify products based on information such as product introductions and pictures. However, due to the strong subjectivity of manual identification, it affects the identification results and cannot ensure the accuracy of identification.

Contents of the invention

Based on this, it is necessary to provide a method and device, a computer storage medium and a device for authenticating a commodity to address the problem of low identification accuracy.

A method for identifying the authenticity of a commodity, comprising the steps of:

Obtain a product authenticity identification model determined by machine learning based on the characteristic information database of preset sample products through machine learning;

Obtain the characteristic information of the product to be tested;

Using the feature information of the commodity to be tested as the input of the commodity authenticity identification model to identify the authenticity of the commodity, and obtain authenticity identification data;

According to the identification data and the data ranges corresponding to each preset authenticity level, the authenticity level of the commodity to be tested is determined.

Also provided is a commodity authenticity identification device, comprising:

The identification model acquisition module is used to obtain a product authenticity identification model determined by machine learning according to the feature information database of preset sample products through a preset machine learning device;

The first characteristic information acquisition module is used to obtain the characteristic information of the product to be tested;

The identification data acquisition module is used to use the characteristic information of the commodity to be tested as the input of the commodity authenticity identification model to identify the authenticity of the commodity, and obtain the authenticity identification data;

The authenticity identification module is used to determine the authenticity level of the product to be tested according to the identification data and the data ranges corresponding to the preset authenticity levels.

A computer storage medium is also provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above method are realized.

A computer device is also provided, including a memory, a processor, and a computer program stored in the memory and operable on the processor, and the above-mentioned method is realized when the processor executes the computer program.

When using the above method and device, computer storage medium and equipment to identify the authenticity of the commodity to be tested, on the one hand, using the commodity authenticity identification model established according to the characteristic information database of the sample commodity through the preset machine learning method, treat The authenticity of the tested product is identified, and the identification data is obtained, so as to realize the accurate identification of the authenticity of the product to be tested. On the other hand, the authenticity identification is carried out through the above process, and there is no need to manually identify the true and false, avoiding the influence of artificial subjectivity on identification As a result, improve the accuracy of commodity identification, and then determine the authenticity level of the commodity to be tested according to the identification data and the corresponding data ranges of the preset authenticity levels, so as to more intuitively know the authenticity of the commodity to be tested. degree of authenticity.

Description of drawings

Fig. 1 is the flow chart of the commodity authenticity identification method of an embodiment;

Fig. 2 is the flow chart of the commodity authenticity identification method of another embodiment;

Fig. 3 is a schematic diagram of modules of a commodity authenticity identification device of an embodiment;

Fig. 4 is a schematic diagram of modules of a product authenticity identification device according to another embodiment.

Detailed ways

Please refer to FIG. 1 , which provides an embodiment of a commodity authenticity identification method, including the following steps S110 to S140.

S110: Acquiring a product authenticity identification model determined by machine learning based on the characteristic information database of preset sample products through machine learning.

Machine learning method is a kind of method that automatically analyzes and obtains laws from known data, and uses the laws to predict unknown data. Machine learning methods correspond to initial prediction models. Different machine learning methods correspond to different initial prediction models. Initial prediction models Corresponding to the input, process parameters and output, the output is determined by the input and process parameters. Machine learning is the process of continuously correcting the process parameters according to the known input data and corresponding output data to obtain the optimal process parameters. The input data is used as the input of the initial prediction model, and the known output data is used as the output of the initial prediction model, and the process parameters in the initial prediction model are continuously updated to obtain the optimal process parameters, thereby obtaining the target prediction model (in this embodiment, the corresponding commodity true and false identification model). It can be understood that the target prediction model obtained after machine learning is the corresponding relationship between the input and the output established by using the optimal process parameters obtained, that is, the above-mentioned law that the known input data and output data satisfy. When the input data predicts its corresponding output, the unknown input data is used as the input of the target prediction model obtained above, and the prediction result can be obtained through the target prediction model, that is, the output.

In this embodiment, firstly, a pre-determined commodity authenticity identification model is obtained, so that the authenticity of the commodity can be subsequently identified through the authenticity identification model. The above authenticity identification model is obtained by machine learning based on the characteristic information database of preset sample commodities. The corresponding input is the characteristic information of the commodity, and the corresponding output is the true and false label, which is established according to the target process parameters. The corresponding relationship between the characteristic information of the commodity and the true and false labels. Specifically, the feature information library of the preset sample products is used as the known data required by the machine learning method, and then machine learning is performed to find the rules that the feature information of the preset sample products in the feature information library of the preset sample products satisfies.

S120: Obtain feature information of the product to be tested.

After the authenticity identification model of the product is obtained, the authenticity identification of the product to be tested can be carried out. First, the characteristic information of the product to be tested needs to be obtained as the input of the product authenticity identification model. In this embodiment, the feature information of a single commodity to be tested has the same number of features and the same feature type as the feature information of a single sample commodity. For example, the sample product includes product A and product B, and the feature information corresponding to product A and product B includes A1 feature and A2 feature. After obtaining the corresponding product authenticity identification model, the product C to be tested is predicted, and the obtained test product The feature information of product C also includes A1 feature and A2 feature, that is, the A1 feature and A2 feature of the product C to be tested, which provide data basis for the subsequent authentication of the product C to be tested through the product authenticity identification model.

S130: Using the feature information of the commodity to be tested as an input of the commodity authenticity identification model to identify the authenticity of the commodity, and obtain authenticity identification data.

After obtaining the authenticity identification model of the commodity and the feature information of the commodity to be tested, the authenticity of the commodity to be tested can be predicted, and the prediction result is the authenticity discrimination data. After establishing the corresponding relationship between the characteristic information of the commodity and the true and false labels according to the target process parameters, the characteristic information of different commodities is used as the input of the authenticity identification model, and the output prediction results may be different. The characteristic information of the commodity to be tested is used as the authenticity of the commodity The input of the identification model is used to identify the authenticity of the commodity, and the authenticity identification data corresponding to the commodity to be tested (data reflecting the authenticity of the commodity to be tested) is obtained.

S140: Determine the authenticity level of the product to be tested according to the identification data and the data ranges corresponding to each preset authenticity level.

Due to the diversity of the obtained authenticity identification data, the authenticity of the product to be tested cannot be accurately judged. The authenticity level is divided in advance. Each authenticity level corresponds to a data range. According to the identification data and each preset authenticity The data ranges corresponding to the false grades determine the true and false grades of the commodities to be tested. Specifically, the preset authenticity level corresponding to the data range to which the identification data belongs is used as the authenticity level of the product to be tested, so that an accurate authenticity identification result of the product can be obtained. And can uniformly and intuitively reflect the authenticity of the goods.

When the authenticity of the commodity to be tested is identified by the above-mentioned commodity authenticity identification method, on the one hand, the commodity authenticity identification model established by the preset machine learning method based on the characteristic information database of the sample commodity is used to identify the authenticity of the commodity to be tested, and the obtained Identify the data, so as to realize the accurate identification of the authenticity of the product to be tested. On the other hand, through the above-mentioned process to identify the authenticity, there is no need to manually identify the authenticity, avoiding the problem of artificial subjectivity affecting the identification results, and improving the accuracy of product identification , and then determine the authenticity level of the product under test according to the identification data and the data ranges corresponding to the preset authenticity levels, so as to know the degree of authenticity of the commodity under test more intuitively. It can provide users with accurate reference information on whether the product to be tested is a fake or inferior product, and improve the user's e-commerce shopping experience, which is of positive significance for promoting the healthy development of the e-commerce industry.

Please refer to Fig. 2, in one of the embodiments, before obtaining the authenticity identification model of commodities determined by machine learning according to the feature information library of preset sample commodities through machine learning methods, further steps are included:

S101: Obtain feature information of each sample commodity and authenticity labels corresponding to each sample commodity.

Before acquiring the authenticity identification model of the above-mentioned commodities, the authenticity identification model of the commodities must first be determined. Authentic and false labels, wherein the authentic and false labels are used to mark the authenticity of the commodity. The authentic and false label library of the commodity is established in advance. The true and false labels corresponding to each sample commodity are obtained from the true and false label library. For example, label 1 can be used to mark the product as true, and label 0 can be used to mark the product as false. For the known sample product a and sample product b, if sample product a is a real product, then the true and false label corresponding to sample product a is the symbol The information that the product is genuine can be 1, and the sample product b is a fake product, then the true and false label corresponding to the sample product b is the information indicating that the product is fake, which can be 0.

Further, each sample product includes a real product and a fake product, and the feature information of each sample product includes the feature information of the real product and the feature information of the fake product, so that the feature information of both the real product and the fake product is covered, which is beneficial to improve The accuracy of the commodity identification model obtained by subsequent machine learning.

S102: According to the characteristic information of each sample commodity and the true and false labels corresponding to each sample commodity, establish a characteristic information database of preset sample commodities.

After obtaining the feature information of each sample product and the corresponding true and false labels of each sample product, they can be combined to establish a feature information database of preset sample products for unified storage and subsequent machine learning. Information and access to true and false labels. The characteristic information database of the preset sample commodities includes characteristic information of each sample commodity and authenticity labels corresponding to each sample commodity.

S103: Using a machine learning method, perform machine learning according to the characteristic information of each sample commodity in the established sample commodity library and the authenticity labels corresponding to each sample commodity, and determine a commodity authenticity identification model.

In a specific example, machine learning methods may include support vectors, neural networks, hidden Markov models, and Bayesian discriminant models, and any one of the above machine learning methods may be used in the process of determining the authenticity of commodities. Machine learning is performed according to the feature information of each sample commodity in the established sample commodity library and the true and false labels corresponding to each sample commodity.

In one embodiment, the manner of obtaining the characteristic information of each sample commodity includes: grabbing the characteristic information of the sample commodity through a web crawler.

A web crawler is a program or script that automatically grabs information on the Internet according to certain rules. In the field of e-commerce, a variety of commodities are available for purchase, and the feature information of sample commodities can be quickly and accurately captured in the network through a web crawler.

In one of the embodiments, the method of obtaining the feature information of the product to be tested includes: obtaining the web page address of the product to be tested; obtaining the web page of the product to be tested according to the web page address, reading the content of the web page of the product to be tested, Obtain feature information of the product to be tested.

When it is necessary to predict the authenticity of the commodity to be tested, the user can input the webpage address of the commodity to be tested to obtain the network address of the commodity to be tested, and then obtain the webpage of the commodity to be tested according to the webpage address. The webpage includes Relevant information of the product to be tested, for example, the introduction of the time on the shelf, product information (such as material, name, origin, etc.), by reading the content of the web page of the product to be tested, the feature information of the product to be tested can be obtained.

In one embodiment, the feature information includes name information, manufacturer information, origin information, evaluation information, price information, introduction information and specification information.

By combining the above-mentioned name information, manufacturer information, origin information, evaluation information, price information, introduction information, and specification information, the characteristics of the corresponding commodity can be accurately characterized, and the above-mentioned characteristic information is used as the input of the authenticity identification model of the commodity to identify the commodity. During authenticity identification, the accuracy of the authenticity prediction result of the commodity authenticity identification model can be improved.

In one of the embodiments, after the determination of the authenticity level of the commodity to be tested, a step is further included: displaying the authenticity level of the commodity to be tested according to a preset display form.

In a specific example, the preset truth level may include an absolute true level, a true level, a general true level, a general false level, a false level and a serious false level. After the authenticity level of the commodity to be tested is determined, it is displayed so as to more intuitively prompt the user about the authenticity of the commodity, that is, to facilitate the user to check the authenticity of the commodity to be tested. In this embodiment, the display is performed in a preset display form, wherein the preset display form includes a text display form and a color picture display form. Regarding the text display form, when displaying the authenticity level of the commodity under test, the corresponding text information is displayed. For example, the authenticity level of the commodity c under test is the true level, and the true level text is displayed correspondingly. For the color picture display form, the true and false levels correspond to the color pictures, and the colors of pictures displayed by different true and false levels are different. For example, the true and false levels range from absolute true to serious false, and the color of the picture can gradually become darker or lighter. For example, The color of the picture corresponding to the true level is light red. Generally, the color of the picture corresponding to the true level is red. If the authenticity level of the product c to be tested is the true level, a red picture will be displayed to remind the user in a more intuitive way.

Please refer to Figure 3, which provides an embodiment of a product authenticity identification device, including:

The identification model acquisition module 310 is used to acquire a product authenticity identification model determined by machine learning based on the feature information database of preset sample products through a preset machine learning device.

The first feature information acquisition module 320 is configured to acquire feature information of the product to be tested.

The identification data acquisition module 330 is used to use the characteristic information of the product to be tested as the input of the product authenticity identification model to perform authenticity identification of the product and obtain authenticity identification data.

The authenticity identification module 340 is used to determine the authenticity level of the product to be tested according to the identification data and the corresponding data ranges of the preset authenticity levels.

When using the above-mentioned commodity authenticity identification device to identify the authenticity of the commodity to be tested, on the one hand, the authenticity of the commodity to be tested is identified by using the commodity authenticity identification model established by the preset machine learning method according to the characteristic information database of the sample commodity, and the obtained Identify the data, so as to realize the accurate identification of the authenticity of the product to be tested. On the other hand, through the above-mentioned process to identify the authenticity, there is no need to manually identify the authenticity, avoiding the problem of artificial subjectivity affecting the identification results, and improving the accuracy of product identification , and then determine the authenticity level of the product under test according to the identification data and the data ranges corresponding to the preset authenticity levels, so as to know the degree of authenticity of the commodity under test more intuitively. It can provide users with accurate reference information on whether the product to be tested is a fake or inferior product, and improve the user's e-commerce shopping experience, which is of positive significance for promoting the healthy development of the e-commerce industry.

Please refer to Fig. 4, in one of the embodiments, the above-mentioned commodity authenticity identification device also includes:

An information acquisition module 301, configured to acquire the feature information of each sample product and the true and false labels corresponding to each sample product;

The characteristic information database building module 302 is used to establish a characteristic information database of preset sample commodities according to the characteristic information of each sample commodity and the true and false labels corresponding to each sample commodity;

The identification model determination module 303 is used to perform machine learning according to the characteristic information of each sample commodity in the established sample commodity library and the authenticity labels corresponding to each sample commodity through a preset machine learning device, so as to determine a commodity authenticity identification model.

In one of the embodiments, the information acquisition module 301 includes:

The second feature information acquisition module is used to capture feature information of sample commodities through a web crawler.

In one of the embodiments, the above-mentioned second feature information acquisition module includes:

The web page address obtaining module is used to obtain the web page address of the product to be tested.

The characteristic information acquisition module of the commodity under test is used to obtain the webpage of the commodity under test according to the webpage address, read the content of the webpage of the commodity under test, and obtain the characteristic information of the commodity under test.

In one embodiment, the feature information includes name information, manufacturer information, origin information, evaluation information, price information, introduction information and specification information.

In one of the embodiments, the above-mentioned commodity authenticity identification device further includes:

The display module is used to display the authenticity level of the commodity to be tested according to a preset display form.

An embodiment of the present invention also provides a computer storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned commodity authenticity identification method are realized.

An embodiment of the present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and operable on the processor. When the processor executes the computer program, the above method for authenticating a commodity is realized.

The technical features in the above-mentioned commodity authentication device, computer storage medium and computer equipment are respectively corresponding to the technical features in the above-mentioned commodity authentication method, and will not be repeated here.

The technical features of the above embodiments can be combined arbitrarily. To make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be It is considered to be within the range described in this specification.

The above examples only express several implementation modes of the present invention, and the description thereof is relatively specific and detailed, but it should not be construed as limiting the scope of the patent for the invention. It should be pointed out that those skilled in the art can make several modifications and improvements without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent for the present invention should be based on the appended claims.

Claims (10)

1. a kind of commercial articles true and false discrimination method, it is characterised in that comprise the following steps：

Acquisition passes through machine learning method, and the commodity that machine learning determination is carried out according to the characteristic information storehouse of default sample commodity are true Vacation differentiates model；

Obtain the characteristic information of commodity to be measured；

Differentiate that the input of model carries out commercial articles true and false discriminating using the characteristic information of the commodity to be measured as the commercial articles true and false, obtain Obtain true and false authentication data；

According to data area corresponding to the authentication data and each default true and false grade difference, the true of the commodity to be measured is determined False grade.

2. commercial articles true and false discrimination method according to claim 1, it is characterised in that the acquisition passes through machine learning side Method, before the step of commercial articles true and false discriminating model of the characteristic information storehouse of default sample commodity progress machine learning determination, Also include step：

Obtain each sample commodity characteristic information and each sample commodity respectively corresponding to true and false label；

According to the characteristic information of each sample commodity and each sample commodity respectively corresponding to true and false label, described in foundation The characteristic information storehouse of default sample commodity；

Pass through the machine learning method, the characteristic information of each sample commodity in the sample commodity storehouse of foundation And true and false label corresponding to each sample commodity difference carries out machine learning, determines that the commercial articles true and false differentiates model.

3. commercial articles true and false discrimination method according to claim 2, it is characterised in that each sample commodity of acquisition The mode of characteristic information includes：

Pass through the characteristic information of sample commodity described in the crawler capturing of networking.

4. commercial articles true and false discrimination method according to claim 1, it is characterised in that the feature letter for obtaining commodity to be measured The mode of breath includes：

Obtain the web page address of the commodity to be measured；

The webpage of the commodity to be measured is obtained according to the web page address, reads the content of the webpage of the commodity to be measured, is obtained The characteristic information of commodity to be measured.

5. the commercial articles true and false discrimination method according to any one in claim 1-4, it is characterised in that the characteristic information Including name information, manufacturer's information, place of production information, evaluation information, pricing information, recommended information and specification information.

6. commercial articles true and false discrimination method according to claim 1, it is characterised in that described to determine the true of the commodity to be measured After false grade, in addition to step：

The true and false grade of the commodity to be measured is shown according to default display format.

A kind of 7. commercial articles true and false identification device, it is characterised in that including：

Differentiate model acquisition module, for obtaining by default machine learning device, according to the characteristic information of default sample commodity The commercial articles true and false that storehouse carries out machine learning determination differentiates model；

Fisrt feature data obtaining module, for obtaining the characteristic information of commodity to be measured；

Authentication data acquisition module, for differentiating the defeated of model using the characteristic information of the commodity to be measured as the commercial articles true and false Enter to carry out commercial articles true and false discriminating, obtain true and false authentication data；

True and false identification module, for according to the authentication data and each default true and false grade respectively corresponding to data area, really The true and false grade of the fixed commodity to be measured.

8. commercial articles true and false identification device according to claim 7, it is characterised in that also include：

Data obtaining module, for obtaining corresponding to characteristic information and each sample commodity difference of each sample commodity True and false label；

Module is established in characteristic information storehouse, for the characteristic information according to each sample commodity and each sample commodity difference Corresponding true and false label, establish the characteristic information storehouse of the default sample commodity；

Differentiate model determining module, for by the default machine learning device, according in the sample commodity storehouse of foundation Each sample commodity characteristic information and each sample commodity respectively corresponding to true and false label carry out machine learning, really The fixed commercial articles true and false differentiates model.

9. a kind of computer-readable storage medium, is stored thereon with computer program, it is characterised in that the computer program is by processor The step of any one methods described in the claims 1-6 is realized during execution.

10. a kind of computer equipment, including memory, processor and it is stored on the memory and can be in the processor The computer program of upper operation, it is characterised in that realize such as claim 1-6 described in the computing device during computer program Method described in middle any one.