Merchant information determination method, device, electronic device and non-volatile storage medium

By extracting the vocabulary features in the merchant description information, obtaining the probability set of main business categories, and automatically determining the main business categories of merchants, solving the problems of high cost and inaccurate manual tagging, and improving the accuracy of merchant information recommendations.

CN111833118BActive Publication Date: 2024-11-05RAJAX NETWORK &TECHNOLOGY (SHANGHAI) CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN201910284931.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2019-04-10
Publication Date
2024-11-05
Estimated Expiration
2039-04-10

AI Technical Summary

Technical Problem

In the field of restaurant takeaway, offline merchants need to manually fill in the main business category when applying for a platform, resulting in high labor and time costs, and inaccurate markings, affecting the accuracy of online merchant information recommendations.

Method used

By obtaining the merchant's description information, extracting vocabulary features, and obtaining the probability set of main business categories based on these characteristics, automatically determine the merchant's main business categories, reduce the cost of manual tagging, and improve the accuracy of tags.

Benefits of technology

Automatic marking of the main business category of merchants is realized, labor costs and time costs are reduced, and the accuracy of marking and merchant information recommendations are improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN111833118B_ABST
    Figure CN111833118B_ABST
Patent Text Reader

Abstract

The embodiments of the present invention relate to the field of communication technology, and disclose a merchant information determination method, device, electronic device and non-volatile storage medium, including: obtaining description information of a target merchant; extracting vocabulary features contained in the description information of the target merchant; obtaining a first main category probability set corresponding to the target merchant according to the vocabulary features, wherein the first main category probability set contains matching probabilities of at least two main categories corresponding to the vocabulary features; determining the main category of the target merchant according to the first main category probability set, so that the main category of the merchant can be automatically marked, saving labor costs and time costs, and improving the accuracy of marking.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of information and communication technology, and in particular to a merchant information determination method, device, electronic device and non-volatile storage medium. Background Art

[0002] In the field of restaurant delivery, the restaurant's main category is a particularly important attribute for identifying the restaurant's business direction, and can often significantly influence users' choices of online ordering. For example, fruit juice drinks, hot pot, and malatang are all relatively common main categories.

[0003] The inventors have discovered that there are at least the following problems in the related technology: offline merchants will manually fill in the main categories when applying for the platform to go online, but with the refined operation of the platform, the main category structure may need to be added or refined. At this time, if the merchants still use the method of manual marking, it will require a lot of labor costs and time costs, and this method of manually marking the main categories will also affect the accuracy of the marking and the accuracy of the online merchant information recommendations. Summary of the invention

[0004] The purpose of the embodiments of the present invention is to provide a merchant information determination method, device, electronic device and non-volatile storage medium, so that the main categories of the merchant can be automatically marked, labor costs and time costs can be saved, and the accuracy of the marking can be improved.

[0005] To solve the above technical problems, an embodiment of the present invention provides a merchant information determination method, which is applied to an electronic device and includes the following steps: obtaining description information of a target merchant; extracting lexical features contained in the description information of the target merchant; obtaining a first main category probability set corresponding to the target merchant based on the lexical features, wherein the first main category probability set includes a matching probability of each main category corresponding to the lexical features; and marking the target merchant by its main category based on the first main category probability set.

[0006] An embodiment of the present invention also provides a merchant information determination device, including: a first acquisition module, used to obtain the description information of the target merchant; an extraction module, used to extract the vocabulary features contained in the description information of the target merchant; a second acquisition module, used to obtain a first main category probability set corresponding to the target merchant based on the vocabulary features, wherein the first main category probability set contains the matching probability of at least two main categories corresponding to the vocabulary features; a main category recording module, used to determine the main category of the target merchant based on the first main category probability set.

[0007] An embodiment of the present invention also provides an electronic device, including a memory and a processor, the memory storing a computer program, and the processor executing the program to: obtain description information of a target merchant; extract vocabulary features contained in the description information of the target merchant; obtain a first main category probability set corresponding to the target merchant based on the vocabulary features, wherein the first main category probability set includes matching probabilities of at least two main categories corresponding to the vocabulary features; and determine the main category of the target merchant based on the first main category probability set.

[0008] The embodiment of the present invention further provides a non-volatile storage medium for storing a computer-readable program, and the computer-readable program is used for a computer to execute the above merchant information determination method.

[0009] Compared with the prior art, the main difference and effect of the embodiments of the present invention are: it is possible to obtain the first main category probability set corresponding to the target merchant based on the vocabulary features contained in the description information of the target merchant, and determine the main category that the merchant needs to mark based on the obtained first main category probability set, so as to automatically realize the marking of the merchant's main category, reduce the labor cost and time cost of manual marking, and improve the accuracy of marking and the accuracy of merchant information recommendation.

[0010] In addition, the lexical features include: the word segmentation results of the target merchant and the segmentation results of the target merchant; extracting the lexical features contained in the description information of the target merchant, including: performing word segmentation processing on the description information of the target merchant to obtain the word segmentation results of the target merchant; performing single-character segmentation processing on the description information of the target merchant to obtain the segmentation results of the target merchant.

[0011] In addition, obtaining a first main category probability set corresponding to the target merchant based on lexical features includes: inputting the word segmentation result of the target merchant into the first determination model to obtain a second main category probability set corresponding to the target merchant, wherein the second main category probability set includes the matching probability of each main category corresponding to the word segmentation result of the target merchant, and the first determination model includes a first mapping relationship between the word segmentation result and the matching probability of the main category; inputting the segmentation result of the target merchant into the second determination model to obtain a third main category probability set corresponding to the target merchant, wherein the third main category probability set includes the matching probability of each main category corresponding to the segmentation result of the target merchant, and the second determination model includes a second mapping relationship between the segmentation result and the matching probability of the main category; and obtaining the first main category probability set based on the second main category probability set and the third main category probability set. In this implementation, corresponding main category probability sets are obtained based on different types of lexical features extracted from the description information of the target merchant, and a first main category probability set is obtained based on the obtained multiple corresponding main category probability sets, thereby making the obtained first main category probability set more accurate.

[0012] In addition, the first main category probability set is obtained according to the second main category probability set and the third main category probability set, including: adding the probabilities of the same type of main categories in the second main category probability set and the third main category probability set to obtain a fourth main category probability set; retaining the probabilities of main categories of different types in the second main category probability set and the third main category probability set to obtain a fifth main category probability set; merging the fourth main category probability set and the fifth main category probability set to obtain the first main category probability set.

[0013] In addition, the main business category of the target merchant is marked according to the first main business category probability set, including: determining at least one main business category corresponding to a probability value less than a preset threshold in the first main business category probability set; and using at least one main business category as the main business category of the target merchant. In this implementation, the probability value in the first main business category probability set reflects the matching similarity between the main business category and the target merchant, and by using at least one main business category corresponding to a probability value less than the preset threshold as the main business category of the target merchant, the main business category of the target merchant determined is more accurate.

[0014] In addition, before obtaining the first main category probability set corresponding to the target merchant according to the lexical features, it also includes: obtaining initial information of the sample merchants, wherein the initial information includes the description information of the sample merchants and the marked main categories, and the number of sample merchants is at least two; performing word segmentation processing on the description information of the sample merchants to obtain the word segmentation results of the sample merchants; performing single-character segmentation processing on the description information of the sample merchants to obtain the segmentation results of the sample merchants; sequentially connecting the main categories and the word segmentation results of the sample merchants to obtain the first sample information; sequentially connecting the main categories and the segmentation results of the sample merchants to obtain the second sample information; obtaining a first determination model according to the first sample information, and obtaining a second determination model according to the second sample information.

[0015] In addition, obtaining a first determination model according to the first sample information and obtaining a second determination model according to the second sample information include: training the first sample information through the fasttext algorithm to obtain the first determination model; training the second sample information through the fasttext algorithm to obtain the second determination model.

[0016] In addition, the word segmentation results of the target merchant are input into the first determination model to obtain the second main category probability set corresponding to the target merchant, including: obtaining the matching probability of each main category corresponding to the word segmentation results of the target merchant according to the word segmentation results of the target merchant and the first determination model; obtaining the second main category probability set according to the matching probability of each main category corresponding to the word segmentation results of the target merchant.

[0017] In addition, the segmentation results of the target merchant are input into the second determination model to obtain the third main category probability set corresponding to the target merchant, including: obtaining the matching probability of each main category corresponding to the segmentation results of the target merchant according to the segmentation results of the target merchant and the second determination model; obtaining the third main category probability set according to the matching probability of each main category corresponding to the segmentation results of the target merchant.

[0018] In addition, after performing word segmentation processing on the description information of the target merchant to obtain the word segmentation result of the target merchant, before inputting the word segmentation result of the target merchant into the first determination model to obtain the second main category probability set corresponding to the target merchant, it also includes: filtering the noise in the word segmentation result of the target merchant, wherein the noise includes punctuation marks, numbers and stop words. In this implementation, by filtering the noise in the word segmentation result of the target merchant, the inaccuracy of the obtained second main category probability set caused by the existence of noise in the word segmentation result of the target merchant can be eliminated.

[0019] In addition, after performing single-character segmentation processing on the description information of the target merchant to obtain the segmentation result of the target merchant, before inputting the segmentation result of the target merchant into the second determination model to obtain the third main category probability set corresponding to the target merchant, it also includes: filtering the noise in the segmentation result of the target merchant, wherein the noise includes punctuation marks, numbers and stop words. In this implementation, by filtering the noise in the segmentation result of the target merchant, it is possible to avoid the inaccuracy of the obtained third main category probability set due to the existence of noise in the segmentation result of the target merchant.

[0020] In addition, the description information of the target merchant includes: the name of the target merchant, the product name of the target merchant and the user evaluation of the target merchant. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 is a flow chart of a method for determining merchant information according to a first embodiment of the present invention;

[0022] Figure 2 is a flowchart of a merchant information determination method in a specific implementation according to the first embodiment of the present invention;

[0023] Figure 3 is a flowchart of a method for determining merchant information according to a second embodiment of the present invention;

[0024] Figure 4 is a schematic diagram of a merchant information determination device according to a third embodiment of the present invention;

[0025] Figure 5 is a schematic structural diagram of an electronic device provided according to a fourth embodiment of the present invention. DETAILED DESCRIPTION

[0026] To make the purpose, technical scheme and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings. However, it will be appreciated by those skilled in the art that in the embodiments of the present invention, many technical details are proposed in order to enable the reader to better understand the present application. However, even without these technical details and various changes and modifications based on the following embodiments, the technical scheme claimed in the present application can be implemented. The division of the following embodiments is for the convenience of description, and should not constitute any limitation on the specific implementation of the present invention. The various embodiments can be combined and referenced with each other without contradiction.

[0027] The first embodiment of the present invention relates to a method for determining merchant information, which is applied to an electronic device. The flowchart of the method can be as follows: Figure 1 As shown, specifically including:

[0028] Step 101, obtaining description information of the target merchant.

[0029] Specifically, in this embodiment, the description information of the target merchant includes the name of the target merchant, the product name of the target merchant, and the user evaluation of the target merchant. For example, the description information of the target merchant includes "The fragrant spicy hot pot is very delicious", where "the fragrant spicy hot pot" represents the name of the target merchant, "the spicy hot pot" represents the product name of the target merchant, and "very delicious" represents the user evaluation of the target merchant. Of course, this embodiment is only an example for explanation, and does not limit the specific content of the description information of the target merchant. In addition, this embodiment only takes the description information of the target merchant as an example to illustrate that one product and user evaluation are included, and in actual applications, the description information of the target merchant can also include multiple products and multiple user evaluations. For example, the description information of the merchant to be calibrated may also include "Fragrant spicy hot pot, spicy hot pot very delicious and tasty", wherein "spicy hot pot" and "spicy hot pot" respectively represent the product names of the target merchant, and "very delicious" and "spicy and tasty" respectively represent user reviews of the merchant to be calibrated. When the description information of the target merchant includes multiple products and multiple user reviews, the method of determining the main category of the target merchant is roughly the same as that of only including one product and user reviews, except that the determination of the main category of the target merchant is more accurate. Therefore, the number of products and the number of user reviews in the description information of the target merchant can be determined according to the actual needs of the user, which will not be repeated in this embodiment.

[0030] It should be noted that the description information of the above-mentioned target merchants can be obtained through collection on the online food delivery platform. Of course, it can also be obtained in other ways. The implementation method of this application does not limit the specific method of obtaining the description information of the target merchants. As long as the description information of the target merchants can be accurately obtained, it is within the scope of protection of this application.

[0031] Step 102: extracting vocabulary features contained in the description information of the target merchant.

[0032] Specifically, in this embodiment, the vocabulary features include the word segmentation results of the target merchant and the segmentation results of the target merchant, and the word segmentation results of the target merchant are obtained by performing word segmentation processing on the description information of the target merchant, and the segmentation results of the target merchant are obtained by performing single-character segmentation processing on the description information of the target merchant.

[0033] For example, the description information of the target merchant includes "Piaoxiang flavor spicy hot pot is delicious", and the segmentation result obtained by performing word segmentation processing on the description information of the target merchant is "Piaoxiang flavor spicy hot pot is delicious", and the segmentation result obtained by performing single-character segmentation processing on the description information of the target merchant is "Piaoxiang flavor spicy hot pot is delicious".

[0034] It should be noted that, in actual applications, if it is determined that there is noise in the word segmentation results of the target merchant or the single-character segmentation results of the target merchant, it is usually necessary to filter the noise in the word segmentation results or the single-character segmentation results, and the noise generally includes punctuation marks, numbers, stop words, etc. Through noise filtering, the accuracy of the main category marked by the description information of the target merchant can be improved.

[0035] Step 103: Acquire a first main category probability set corresponding to the target merchant according to the vocabulary features.

[0036] Specifically, in this embodiment, the first main category probability set includes the matching probability of at least two main categories corresponding to the vocabulary features. The main categories in the first main category probability set refer to the categories of the main business commodities of the merchant, for example, they may include: fruit juice drinks, hot pot, spicy hot pot, Sichuan cuisine, Shandong cuisine, desserts, etc., and the main categories in the first main category probability set are set according to the actual needs of the user.

[0037] In a specific implementation, the specific process of this step is as follows Figure 2 As shown, it includes the following sub-steps:

[0038] Step 1031: Input the word segmentation result of the target merchant into the first determination model to obtain a second main category probability set corresponding to the target merchant.

[0039] Specifically, in this embodiment, the first determination model includes a first mapping relationship between the word segmentation results and the matching probabilities of the main categories. According to the word segmentation results of the target merchant and the first determination model, the matching probability of each main category corresponding to the word segmentation results of the target merchant is obtained, and according to the matching probability of each main category corresponding to the word segmentation results of the target merchant, a second main category probability set is obtained. Therefore, the second main category probability set contains the matching probability of each main category corresponding to the word segmentation results of the target merchant.

[0040] For example, if the word segmentation result of the target merchant is determined to be "the spicy hot pot with a fragrant flavor is delicious", by inputting the word segmentation result of the target merchant into the first determination model, and using the first mapping relationship between the word segmentation result and the matching probability of the main category in the first determination model, the matching probability of each two main categories corresponding to the word segmentation result of the target merchant can be obtained as "hot pot 65% spicy hot pot 20% Sichuan cuisine 15%", so the second main category probability set corresponding to the target merchant is {hot pot 65% spicy hot pot 20% Sichuan cuisine 15%}. Of course, this embodiment is only an example for explanation, and the method for obtaining the second main category probability set for the word segmentation results of other types of target merchants is roughly the same, so it will not be repeated in this embodiment.

[0041] It should be noted that in order to ensure the accuracy of the determination of the merchant's main categories, the first standard threshold can be set according to actual needs, and the main categories in the second main category probability set that are lower than the first standard threshold are deleted. For example, the first standard threshold can be set to 5%, and the main categories in the first main category probability set and the second main category probability set that are lower than 5% are deleted. Of course, this embodiment is only an example for illustration, and the specific value of the first standard threshold is not limited in the embodiment of this application.

[0042] Step 1032: Input the segmentation result of the target merchant into the second determination model to obtain the third main category probability set corresponding to the target merchant.

[0043] Specifically, in this embodiment, the second determination model includes a second mapping relationship between the segmentation results and the matching probabilities of the main categories. Based on the segmentation results of the target merchant and the second determination model, the matching probability of each main category corresponding to the segmentation results of the target merchant is obtained, and based on the matching probability of each main category corresponding to the segmentation results of the target merchant, a third main category probability set is obtained. Therefore, the third main category probability set contains the matching probability of each main category corresponding to the segmentation results of the target merchant.

[0044] For example, if the segmentation result of the target merchant is determined to be "the spicy hot pot with a fragrant flavor is delicious", by inputting the segmentation result of the target merchant into the second determination model, and using the second mapping relationship between the segmentation result and the matching probability of the main category in the second determination model, the matching probability of each main category corresponding to the segmentation result of the target merchant can be obtained as "hot pot 80% spicy hot pot 11% Beijing cuisine 9%", so that the third main category probability set corresponding to the target merchant is {hot pot 80% spicy hot pot 11% Beijing cuisine 9%}. Of course, this embodiment is only an example for explanation, and the method for obtaining the third main category probability set for the segmentation results of other types of target merchants is roughly the same, so it will not be repeated in this embodiment.

[0045] It should be noted that in order to ensure the accuracy of determining the merchant's main categories, a second standard threshold can be set according to actual needs, and the main categories in the third main category probability set that are lower than the second standard threshold can be deleted. The specific value of the second standard threshold can be the same as or different from the first standard threshold. The specific value of the second standard is not limited in the implementation mode of the present application.

[0046] Step 1033: Obtain a first main category probability set according to the second main category probability set and the third main category probability set.

[0047] Specifically, in this embodiment, the probabilities of the same type of main categories in the second main category probability set and the third main category probability set are added to obtain a fourth main category probability set. The probabilities of different types of main categories in the second main category probability set and the third main category probability set are retained to obtain a fifth main category probability set, and the fourth main category probability set and the fifth main category probability set are merged to obtain a first main category probability set.

[0048] For example, for the description information of the target merchant "The spicy hot pot with a fragrant flavor is delicious", if the second main category probability set obtained is {hot pot 65% spicy hot pot 20% Sichuan cuisine 15%}, and the third main category probability set is {hot pot 80% spicy hot pot 11% Beijing cuisine 9%}, then it is determined that the main categories of the same type in the second main category probability set and the third main category probability set are "hot pot" and "spicy hot pot" respectively, and the probabilities of the same type of main categories are added to obtain "hot pot 145% spicy hot pot 31%", so the fourth main category probability set is {hot pot 65% spicy hot pot 20% Sichuan cuisine 15%}. Hot pot 145% Malatang 31%}; and determine that the different types of main categories in the second main category probability set and the third main category probability set are "Sichuan cuisine" and "Beijing cuisine", respectively, retain the probabilities of different types of main categories, and obtain "Sichuan cuisine 15% Beijing cuisine 9%", so the fifth main category probability set is {Sichuan cuisine 15% Beijing cuisine 9%}, and by merging the fourth main category probability set with the fifth main category probability set, the obtained first main category probability set is {Hot pot 145% Malatang 31% Sichuan cuisine 15% Beijing cuisine 9%}.

[0049] Step 104: Mark the target merchant by its main category according to the first main category probability set.

[0050] Specifically, in this embodiment, at least one main category corresponding to the probability value greater than the preset threshold in the first main category probability set is determined, and at least one main category is used as the main category of the target merchant. For example, for the description information of the target merchant "The spicy hot pot with fragrant taste is delicious", if the first main category probability set obtained is {hot pot 145%, spicy hot pot 31%, Sichuan cuisine 15%, Beijing cuisine 9%}, and the preset threshold is 100%, and the preset threshold in this embodiment can be set by the user according to actual needs, of course, it can also be a value set by statistics based on historical calibration records, and the specific method of obtaining the preset threshold is not limited in this embodiment. Because the main category corresponding to the probability value greater than the preset threshold in the first main category probability set is "hot pot", the main category "hot pot" can be used as the main category of the target merchant "fragrant taste". Of course, the present embodiment uses a main category that exceeds a preset threshold as an example for illustration. There may also be multiple main categories whose probability values ​​exceed the preset threshold. In this case, the multiple main categories obtained can be determined as the main categories corresponding to the target merchant, and it can be indicated that the target merchant operates multiple main categories at the same time.

[0051] Compared with the prior art, the merchant information determination method provided in this embodiment can obtain the first main category probability set corresponding to the target merchant based on the vocabulary features contained in the description information of the target merchant, and determine the main category that the merchant needs to mark based on the obtained first main category probability set, thereby automatically marking the merchant's main category, reducing the labor cost and time cost of manual marking, and improving the accuracy of marking and the accuracy of merchant information recommendation.

[0052] The second embodiment of the present invention relates to a merchant information determination method. This embodiment is further improved on the basis of the first embodiment. The specific improvement is: before inputting the segmentation result of the target merchant into the first determination model and inputting the segmentation result of the target merchant into the second determination model, a step of obtaining the first determination model and the second determination model according to the description information of the sample merchant is added. The process of the merchant information determination method in this embodiment is as follows: Figure 3 Specifically, in this embodiment, steps 201 to 207 are included, wherein steps 201 to 202 are substantially the same as steps 101 to 102 in the first embodiment, steps 204 to 206 are substantially the same as steps 1031 to 1033 in the first embodiment, and step 207 is substantially the same as step 104 in the first embodiment, which will not be described in detail here, and the differences will be mainly introduced below. For technical details not described in detail in this embodiment, please refer to the merchant information determination method provided in the first embodiment, which will not be described in detail here.

[0053] After step 201 to step 202, step 203 is executed.

[0054] Step 203: Acquire a first determination model and a second determination model according to the description information of the sample merchant.

[0055] Specifically, in this embodiment, the method of obtaining the first determination model and the second determination model can be specifically: obtaining the initial information of the sample merchant, and the initial information includes the description information of the sample merchant and the main category of the sample merchant marked, and the number of sample merchants is at least two, performing word segmentation processing on the description information of the sample merchant to obtain the word segmentation result of the sample merchant, and performing single-character segmentation processing on the description information of the sample merchant to obtain the segmentation result of the sample merchant. The main category and the word segmentation result of the sample merchant are sequentially connected to obtain the first sample information, the main category and the segmentation result of the sample merchant are sequentially connected to obtain the second sample information, the first determination model is obtained according to the first sample information, and the second determination model is obtained according to the second sample information.

[0056] The method of obtaining the first determination model according to the first sample information and the method of obtaining the second determination model according to the second sample information may specifically be: training the first sample information by the fasttext algorithm to obtain the first determination model. In addition, training the second sample information by the fasttext algorithm to obtain the second determination model. Although the word segmentation results in the first sample information have been manually labeled with main categories, the first determination model itself does not have the function of identifying the main categories corresponding to the word segmentation results without sufficient samples for learning and training. Therefore, it is necessary to use the fasttext algorithm to undergo continuous training based on the first sample information. At the same time, in order to ensure the accuracy of the first determination model training, the number of first sample information must be sufficient, and the first sample information must cover as many or even all main categories in the field as possible. At the same time, an error threshold must be set. When the training error result of the fasttext algorithm is infinitely close to the error threshold and remains unchanged, it can be determined that the training is completed. Due to limitations such as the number of first samples and training time, each word segmentation result does not necessarily match its labeled main category 100%. At this time, the first determination model will obtain the matching probability of the word segmentation result with multiple main categories, but its probability with the main category labeled by itself is the largest, and the sum of the matching probabilities of multiple main categories corresponding to one word segmentation result is 1.

[0057] For example, there are multiple sample merchants in the description information of the sample merchants, and four of them, "Piaoxiang Flavor Spicy Hot Pot is Delicious", "Xiangyuan Tianxia Xiangxi Overlord Frog Tastes Too Spicy", "Mala Taste Malatang is Delicious" and "Duck World Roast Duck is Crispy", are used as examples for explanation, and the main category is marked in the description information of each sample merchant to obtain the initial information "Hot Pot Piaoxiang Flavor Spicy Hot Pot is Delicious", "Sichuan Cuisine Xiangyuan Tianxia Xiangxi Overlord Frog Tastes Too Spicy", "Malatang Malatang is Delicious" and "Beijing Cuisine Duck World Roast Duck is Crispy". The description information of the sample merchants in the initial information is segmented to obtain the segmentation results of the sample merchants: "Piaoxiang Taste Spicy Hot Pot is Delicious", "Xiangyuan Tianxia Xiangxi Overlord Frog Tastes Too Spicy", "Mala Taste Spicy Hot Pot is Delicious" and "Duck World Roast Duck is Crispy", and the description information of the sample merchants in the initial information is segmented by single characters to obtain the segmentation results of the sample merchants: "Piaoxiang Taste Spicy Hot Pot is Delicious", "Xiangyuan Tianxia Xiangxi Overlord Frog Tastes Too Spicy", "Mala Taste Spicy Hot Pot is Delicious" and "Duck World Roast Duck is Crispy". The main categories in the initial information and the segmentation results of the sample merchants are sequentially connected to obtain the first sample information: "Hot Pot Piaoxiang Taste Spicy Hot Pot is Delicious", "Sichuan Cuisine Xiangyuan Tianxia Xiangxi Overlord Frog Tastes Too Spicy", "Mala Tang Spicy Hot Pot is Delicious" and "Beijing Cuisine Duck World Roast Duck is Crispy". Therefore, the first sample information in this embodiment is a data set, which includes the segmentation results of multiple sample merchants and their marked main categories. The main categories in the initial information are sequentially connected with the segmentation results of the sample merchants to obtain the second sample information: "Hotpot Piaoxiang Taste Spicy Hotpot is Delicious", "Sichuan Cuisine Xiangyuan Tianxia Xiangxi Overlord Frog Tastes Too Spicy", "Malatang Spicy Taste Malatang is Delicious" and "Beijing Cuisine Duck World Roast Duck is Crispy". Therefore, the second sample information in this embodiment is also a data set, which includes the segmentation results of multiple sample merchants and their marked main categories. Although four examples are used for illustration in this embodiment, the number of sample merchants in actual applications is often very large.

[0058] For example, the first sample information "hot pot smells spicy hot pot is delicious", "Sichuan cuisine Xiangyuan Tianxia Xiangxi Overlord Frog tastes too spicy", "spicy spicy hot pot is delicious" and "Beijing cuisine Duck World Roast Duck is crispy" are trained by the fasttext algorithm, and the error threshold is set to 0.2. Although it is hoped that the training error is 0 under ideal conditions, the model may not be able to fully achieve this due to limitations such as the number of samples and training time. Therefore, the specific value of the error threshold can be set based on experience and actual needs. Therefore, the specific value of the error threshold is not limited in this implementation. When the fasttext algorithm is used for multiple trainings, it is finally determined that the training error is 0.225 and remains unchanged. At this time, although the training error has not reached the error threshold, it is infinitely close to it. At this time, the training stops. The first determination model will obtain the corresponding relationship between the word segmentation result and the matching probability of the main category, and this corresponding relationship is called the first mapping relationship. For example, the first mapping relationship includes: the word segmentation result "Piaoxiang flavor spicy hot pot is delicious" corresponds to the matching probability: hot pot 65%, malatang 20% ​​and Sichuan cuisine 15%; "Xiangyuan Tianxia Xiangxi Overlord Frog tastes too spicy" corresponds to the matching probability: Beijing cuisine 20% and Sichuan cuisine 80%; "Spicy taste malatang is delicious" corresponds to the matching probability: malatang 80%, hot pot 15% and Beijing cuisine 5%; "Duck World Roast Duck is crispy" corresponds to the matching probability: Beijing cuisine 85% and Hunan cuisine 15%. And in this embodiment, the training error obtained in the first determination model refers to the average value of the matching probability errors of all word segmentation results in the first mapping relationship obtained after training. For example, when the first word segmentation "the fragrant spicy hot pot is delicious" is input into the model, the actual main category corresponding to it is "hot pot", and the actual matching probability of "hot pot" actually obtained after training is 65%, so 1-65%=35%, which means that the error of the matching probability of the first word segmentation result "the fragrant spicy hot pot is delicious" in the first mapping relationship is 0.35. Similarly, it can be obtained that the error of the matching probability of "Xiangyuan Tianxia Xiangxi Overlord Frog tastes too spicy" is 0. .2; the error of the matching probability of "spicy spicy hot pot is delicious" is 0.2; the error of the matching probability of "duck world roast duck is crispy" is 0.15, so the average value of the matching probability error of all word segmentation results in the first mapping relationship is (0.35+0.2+0.2+0.15) / 4=0.225, and after multiple trainings, the training error always remains at 0.225. Although the set error threshold is 0.2 at this time, 0.225 is infinitely close to 0.2 and remains unchanged. Due to the limitations of the first sample quantity, training time, etc., it can be determined that the training is over at this time, and 0.225 is used as the final training error result.

[0059] Therefore, in this embodiment, the first sample information is trained by the fasttext algorithm to obtain the first mapping relationship between the word segmentation result of each sample merchant in the first sample information and the matching probability of the main category. When the first mapping relationship is determined, after the word segmentation result of the target merchant is input into the first determination model, the first mapping relationship can be used to directly output the matching probability of the main category of the target merchant. Similarly, the method of obtaining the second determination model is roughly the same as the method of obtaining the first determination model, and will not be repeated in the embodiment of this application.

[0060] It should be noted that, in actual applications, the principle of the first determination model obtaining the matching probability corresponding to the target merchant's word segmentation result is that when the first determination model determines that the similarity between the target merchant's word segmentation result and the word segmentation result of a sample merchant exceeds a preset threshold, the matching probability of the main category corresponding to the word segmentation result of the sample merchant is used as the matching probability of the main category corresponding to the word segmentation result of the target merchant, thereby obtaining the first main category probability set. Similarly, the way in which the second determination model obtains the second main category probability set corresponding to the segmentation result of the target merchant is roughly the same, so it will not be repeated in this embodiment.

[0061] After step 203, steps 204 to 207 are executed.

[0062] Compared with the prior art, the merchant information determination method provided in this embodiment can obtain the first main category probability set corresponding to the target merchant according to the vocabulary features contained in the description information of the target merchant, and determine the main category that the merchant needs to mark according to the obtained first main category probability set, so as to automatically realize the marking of the merchant's main category, reduce the labor cost and time cost of manual marking, and improve the accuracy of marking and the accuracy of merchant information recommendation. And by obtaining the first determination model and the second determination model through the description information of the sample merchant, the accuracy of the main category marking by the first determination model and the second determination model can be guaranteed.

[0063] The step division of the above methods is only for the purpose of clear description. When implemented, they can be combined into one step or some steps can be split and decomposed into multiple steps. As long as they include the same logical relationship, they are all within the scope of protection of this patent; adding insignificant modifications to the algorithm or process or introducing insignificant designs without changing the core design of the algorithm and process are all within the scope of protection of this patent.

[0064] A third embodiment of the present invention relates to a merchant information determination device, such as Figure 4 As shown, the device includes: a first acquisition module 31, an extraction module 32, a second acquisition module 33 and a main category recording module 34.

[0065] The first acquisition module 31 is used to acquire the description information of the target merchant.

[0066] The extraction module 32 is used to extract the vocabulary features contained in the description information of the target merchant.

[0067] The second acquisition module 33 is used to acquire a first main category probability set corresponding to the target merchant according to the vocabulary feature.

[0068] The main category determination module 34 is used to determine the main category of the target merchant according to the first main category probability set.

[0069] It is not difficult to find that this embodiment is a device embodiment corresponding to the first embodiment, and this embodiment can be implemented in conjunction with the first embodiment. The relevant technical details mentioned in the first embodiment are still valid in this embodiment, and in order to reduce repetition, they are not repeated here. Accordingly, the relevant technical details mentioned in this embodiment can also be applied in the first embodiment.

[0070] A fourth embodiment of the present invention relates to an electronic device, such as Figure 5 As shown, the electronic device includes: at least one processor 501; and a memory 502 in communication with the at least one processor 501, wherein the memory 502 stores instructions that can be executed by the at least one processor 501, and the instructions are executed by the at least one processor 501 to implement:

[0071] Obtain description information of a target merchant; extract lexical features contained in the description information of the target merchant; obtain a first main category probability set corresponding to the target merchant based on the lexical features, wherein the first main category probability set includes a matching probability of each main category corresponding to the lexical features; and mark the target merchant by its main category based on the first main category probability set.

[0072] Specifically, the electronic device includes: one or more processors 501 and a memory 502, Figure 5 A processor 501 is taken as an example. The processor 501 and the memory 502 may be connected via a bus or other means. Figure 5 In the example, the bus connection is used. The memory 502 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules. The processor 501 executes various functional applications and data processing of the device by running the non-volatile software programs, instructions and modules stored in the memory 502, that is, realizing the above-mentioned merchant information determination method.

[0073] The memory 502 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and applications required for at least one function; the data storage area may store a list of options, etc. In addition, the memory 502 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 502 may optionally include a memory 502 remotely arranged relative to the processor 501, and these remote memories 502 may be connected to an external device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0074] One or more modules are stored in the memory 502 , and when executed by one or more processors 501 , the merchant information determination method in any of the above method implementations is executed.

[0075] The above-mentioned product can execute the method provided in the implementation mode of this application, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not fully described in this implementation mode, please refer to the merchant information determination method provided in the implementation mode of this application.

[0076] A fifth embodiment of the present application relates to a non-volatile storage medium, in which a computer program is stored. The computer-readable program is used for a computer to execute the merchant information determination method involved in the first or second embodiment of the present invention.

[0077] Those skilled in the art can understand that all or part of the steps in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a program, and the program is stored in a storage medium, including several instructions to enable a device (which can be a single-chip microcomputer, chip, etc.) or a processor to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0078] Those skilled in the art will appreciate that the above-mentioned embodiments are specific examples for implementing the present invention, and in actual applications, various changes may be made thereto in form and detail without departing from the spirit and scope of the present invention.

[0079] The present application embodiment discloses A1. A method for determining merchant information, comprising:

[0080] Get the description information of the target merchant;

[0081] Extracting vocabulary features contained in the description information of the target merchant;

[0082] Acquire a first main category probability set corresponding to the target merchant according to the lexical feature, wherein the first main category probability set includes matching probabilities of at least two main categories corresponding to the lexical feature;

[0083] The main category of the target merchant is determined according to the first main category probability set.

[0084] A2. The merchant information determination method according to A1, wherein the vocabulary features include: the word segmentation result of the target merchant and the segmentation result of the target merchant;

[0085] The extracting of vocabulary features contained in the description information of the target merchant includes:

[0086] Performing word segmentation processing on the description information of the target merchant to obtain a word segmentation result of the target merchant;

[0087] The description information of the target merchant is segmented into single characters to obtain a segmentation result of the target merchant.

[0088] A3. According to the merchant information determination method described in A2, the step of obtaining a first main category probability set corresponding to the target merchant based on the vocabulary feature includes:

[0089] Inputting the word segmentation result of the target merchant into the first determination model to obtain a second main category probability set corresponding to the target merchant, wherein the second main category probability set includes the matching probability of each main category corresponding to the word segmentation result of the target merchant, and the first determination model includes a first mapping relationship between the word segmentation result and the matching probability of the main category;

[0090] Inputting the segmentation result of the target merchant into the second determination model to obtain a third main category probability set corresponding to the target merchant, wherein the third main category probability set includes the matching probability of each main category corresponding to the segmentation result of the target merchant, and the second determination model includes a second mapping relationship between the segmentation result and the matching probability of the main category;

[0091] The first main category probability set is obtained according to the second main category probability set and the third main category probability set.

[0092] A4. According to the merchant information determination method described in A3, obtaining the first main category probability set according to the second main category probability set and the third main category probability set includes:

[0093] Adding the probabilities of the same type of main categories in the second main category probability set and the third main category probability set to obtain a fourth main category probability set;

[0094] retaining the probabilities of different types of main categories in the second main category probability set and the third main category probability set to obtain a fifth main category probability set;

[0095] The fourth main category probability set and the fifth main category probability set are combined to obtain the first main category probability set.

[0096] A5. According to the merchant information determination method described in A1, determining the main category of the target merchant according to the first main category probability set includes:

[0097] Determine at least one main category corresponding to a probability value greater than a preset threshold in the first main category probability set;

[0098] The at least one main category is used as the main category of the target merchant.

[0099] A6. The merchant information determination method according to A3, before obtaining the first main category probability set corresponding to the target merchant according to the vocabulary feature, further comprising:

[0100] Acquire initial information of sample merchants, wherein the initial information includes description information of the sample merchants and marked main categories, and the number of the sample merchants is at least two;

[0101] Performing word segmentation processing on the description information of the sample merchant to obtain the word segmentation result of the sample merchant;

[0102] Performing single-character segmentation processing on the description information of the sample merchant to obtain a segmentation result of the sample merchant;

[0103] Sequentially connect the main business category and the word segmentation results of the sample merchants to obtain first sample information;

[0104] Sequentially connect the segmentation results of the main business category and the sample merchants to obtain second sample information;

[0105] The first determination model is obtained according to the first sample information, and the second determination model is obtained according to the second sample information.

[0106] A7. According to the merchant information determination method described in A6, the obtaining of the first determination model according to the first sample information and the obtaining of the second determination model according to the second sample information include:

[0107] Training the first sample information by using the fasttext algorithm to obtain the first determination model;

[0108] The second sample information is trained by the fasttext algorithm to obtain the second determination model.

[0109] A8. The merchant information determination method according to A7, wherein the step of inputting the word segmentation result of the target merchant into the first determination model to obtain a second main category probability set corresponding to the target merchant comprises:

[0110] According to the word segmentation result of the target merchant and the first determination model, obtaining a matching probability of each main category corresponding to the word segmentation result of the target merchant;

[0111] The second main category probability set is obtained according to the matching probability of each main category corresponding to the word segmentation result of the target merchant.

[0112] A9. According to the merchant information determination method described in A7, the step of inputting the segmentation result of the target merchant into the second determination model to obtain a third main category probability set corresponding to the target merchant includes:

[0113] According to the segmentation result of the target merchant and the second determination model, obtaining a matching probability of each main category corresponding to the segmentation result of the target merchant;

[0114] The third main category probability set is obtained according to the matching probability of each main category corresponding to the segmentation result of the target merchant.

[0115] A10. The merchant information determination method according to A3, after performing word segmentation processing on the description information of the target merchant to obtain the word segmentation result of the target merchant, and before inputting the word segmentation result of the target merchant into the first determination model to obtain the second main category probability set corresponding to the target merchant, further comprising:

[0116] The noise in the word segmentation result of the target merchant is filtered, wherein the noise includes punctuation marks, numbers and stop words.

[0117] A11. The merchant information determination method according to A3, after performing single-character segmentation processing on the description information of the target merchant to obtain the segmentation result of the target merchant, and before inputting the segmentation result of the target merchant into the second determination model to obtain the third main category probability set corresponding to the target merchant, further comprising:

[0118] The noise in the segmentation result of the target merchant is filtered, wherein the noise includes punctuation marks, numbers and stop words.

[0119] A12. According to the merchant information determination method described in A1, the description information of the target merchant includes: the name of the target merchant, the product name of the target merchant and the user evaluation of the target merchant.

[0120] The present application embodiment discloses B1. a merchant information determination device, comprising:

[0121] A first acquisition module is used to acquire description information of a target merchant;

[0122] An extraction module, used to extract vocabulary features contained in the description information of the target merchant;

[0123] A second acquisition module is used to acquire a first main category probability set corresponding to the target merchant according to the lexical feature, wherein the first main category probability set includes matching probabilities of at least two main categories corresponding to the lexical feature;

[0124] A main category determination module is used to determine the main category of the target merchant according to the first main category probability set.

[0125] The present application embodiment discloses C1. an electronic device, comprising:

[0126] Memory and processor. Memory stores computer programs. Processor executes programs when it runs:

[0127] Get the description information of the target merchant;

[0128] Extracting vocabulary features contained in the description information of the target merchant;

[0129] Acquire a first main category probability set corresponding to the target merchant according to the lexical feature, wherein the first main category probability set includes matching probabilities of at least two main categories corresponding to the lexical feature;

[0130] The main category of the target merchant is determined according to the first main category probability set.

[0131] C2. The electronic device according to C1, wherein the vocabulary features include: a word segmentation result of the target merchant and a segmentation result of the target merchant;

[0132] The extracting of vocabulary features contained in the description information of the target merchant includes:

[0133] Performing word segmentation processing on the description information of the target merchant to obtain a word segmentation result of the target merchant;

[0134] The description information of the target merchant is segmented into single characters to obtain a segmentation result of the target merchant.

[0135] C3. The electronic device according to C2, wherein the acquiring, according to the vocabulary feature, a first main category probability set corresponding to the target merchant comprises:

[0136] Inputting the word segmentation result of the target merchant into the first determination model to obtain a second main category probability set corresponding to the target merchant, wherein the second main category probability set includes the matching probability of each main category corresponding to the word segmentation result of the target merchant, and the first determination model includes a first mapping relationship between the word segmentation result and the matching probability of the main category;

[0137] Inputting the segmentation result of the target merchant into the second determination model to obtain a third main category probability set corresponding to the target merchant, wherein the third main category probability set includes the matching probability of each main category corresponding to the segmentation result of the target merchant, and the second determination model includes a second mapping relationship between the segmentation result and the matching probability of the main category;

[0138] The first main category probability set is obtained according to the second main category probability set and the third main category probability set.

[0139] C4. The electronic device according to C3, wherein obtaining the first main category probability set according to the second main category probability set and the third main category probability set comprises:

[0140] Adding the probabilities of the same type of main categories in the second main category probability set and the third main category probability set to obtain a fourth main category probability set;

[0141] retaining the probabilities of different types of main categories in the second main category probability set and the third main category probability set to obtain a fifth main category probability set;

[0142] The fourth main category probability set and the fifth main category probability set are combined to obtain the first main category probability set.

[0143] C5. The electronic device according to C1, wherein determining the main category of the target merchant according to the first main category probability set comprises:

[0144] Determine at least one main category corresponding to a probability value greater than a preset threshold in the first main category probability set;

[0145] The at least one main category is used as the main category of the target merchant.

[0146] C6. The electronic device according to C3, before acquiring the first main category probability set corresponding to the target merchant according to the vocabulary feature, further comprising:

[0147] Acquire initial information of sample merchants, wherein the initial information includes description information of the sample merchants and marked main categories, and the number of the sample merchants is at least two;

[0148] Performing word segmentation processing on the description information of the sample merchant to obtain the word segmentation result of the sample merchant;

[0149] Performing single-character segmentation processing on the description information of the sample merchant to obtain a segmentation result of the sample merchant;

[0150] Sequentially connect the main business category and the word segmentation results of the sample merchants to obtain first sample information;

[0151] Sequentially connect the segmentation results of the main business category and the sample merchants to obtain second sample information;

[0152] The first determination model is obtained according to the first sample information, and the second determination model is obtained according to the second sample information.

[0153] C7. The electronic device according to C6, wherein obtaining the first determination model according to the first sample information and obtaining the second determination model according to the second sample information comprises:

[0154] Training the first sample information by using the fasttext algorithm to obtain the first determination model;

[0155] The second sample information is trained by the fasttext algorithm to obtain the second determination model.

[0156] C8. The electronic device according to C7, wherein the step of inputting the word segmentation result of the target merchant into the first determination model to obtain a second main category probability set corresponding to the target merchant comprises:

[0157] According to the word segmentation result of the target merchant and the first determination model, obtaining a matching probability of each main category corresponding to the word segmentation result of the target merchant;

[0158] The second main category probability set is obtained according to the matching probability of each main category corresponding to the word segmentation result of the target merchant.

[0159] C9. The electronic device according to C7, wherein the step of inputting the segmentation result of the target merchant into a second determination model to obtain a third main category probability set corresponding to the target merchant comprises:

[0160] According to the segmentation result of the target merchant and the second determination model, obtaining a matching probability of each main category corresponding to the segmentation result of the target merchant;

[0161] The third main category probability set is obtained according to the matching probability of each main category corresponding to the segmentation result of the target merchant.

[0162] C10. The electronic device according to C3, after performing word segmentation processing on the description information of the target merchant to obtain the word segmentation result of the target merchant, and before inputting the word segmentation result of the target merchant into the first determination model to obtain the second main category probability set corresponding to the target merchant, further comprising:

[0163] The noise in the word segmentation result of the target merchant is filtered, wherein the noise includes punctuation marks, numbers and stop words.

[0164] C11. The electronic device according to C3, after performing single-character segmentation processing on the description information of the target merchant to obtain the segmentation result of the target merchant, and before inputting the segmentation result of the target merchant into the second determination model to obtain the third main category probability set corresponding to the target merchant, further comprising:

[0165] The noise in the segmentation result of the target merchant is filtered, wherein the noise includes punctuation marks, numbers and stop words.

[0166] C12. The electronic device according to C1, wherein the description information of the target merchant includes: the name of the target merchant, the product name of the target merchant, and a user evaluation of the target merchant.

[0167] The embodiment of the present application discloses D1. A non-volatile storage medium for storing a computer-readable program, wherein the computer-readable program is used for a computer to execute the merchant information determination method as described in any one of A1 to A12.

Claims

1. A method for determining merchant information, characterized in that: include: Get the description information of the target merchant; Extracting lexical features contained in the description information of the target merchant, wherein the lexical features include a word segmentation result of the target merchant and a segmentation result of the target merchant; Acquire a first main category probability set corresponding to the target merchant according to the lexical feature, wherein the first main category probability set includes matching probabilities of at least two main categories corresponding to the lexical feature; Determining the main category of the target merchant according to the first main category probability set; The step of acquiring a first main category probability set corresponding to the target merchant according to the vocabulary feature includes: Inputting the word segmentation result of the target merchant into the first determination model to obtain a second main category probability set corresponding to the target merchant, wherein the second main category probability set includes the matching probability of each main category corresponding to the word segmentation result of the target merchant, and the first determination model includes a first mapping relationship between the word segmentation result and the matching probability of the main category; Inputting the segmentation result of the target merchant into the second determination model to obtain a third main category probability set corresponding to the target merchant, wherein the third main category probability set includes the matching probability of each main category corresponding to the segmentation result of the target merchant, and the second determination model includes a second mapping relationship between the segmentation result and the matching probability of the main category; The first main category probability set is obtained according to the second main category probability set and the third main category probability set.

2. The merchant information determination method according to claim 1, characterized in that: The extracting of vocabulary features contained in the description information of the target merchant includes: Performing word segmentation processing on the description information of the target merchant to obtain a word segmentation result of the target merchant; The description information of the target merchant is segmented into single characters to obtain a segmentation result of the target merchant.

3. The merchant information determination method according to claim 1, characterized in that: The obtaining the first main category probability set according to the second main category probability set and the third main category probability set includes: Adding the probabilities of the same type of main categories in the second main category probability set and the third main category probability set to obtain a fourth main category probability set; retaining the probabilities of different types of main categories in the second main category probability set and the third main category probability set to obtain a fifth main category probability set; The fourth main category probability set and the fifth main category probability set are combined to obtain the first main category probability set.

4. The merchant information determination method according to claim 1, characterized in that: The determining the main category of the target merchant according to the first main category probability set includes: Determine at least one main category corresponding to a probability value greater than a preset threshold in the first main category probability set; The at least one main category is used as the main category of the target merchant.

5. The merchant information determination method according to claim 1, characterized in that: Before acquiring the first main category probability set corresponding to the target merchant according to the vocabulary feature, the method further includes: Acquire initial information of sample merchants, wherein the initial information includes description information of the sample merchants and marked main categories, and the number of the sample merchants is at least two; Performing word segmentation processing on the description information of the sample merchant to obtain the word segmentation result of the sample merchant; Performing single-character segmentation processing on the description information of the sample merchant to obtain a segmentation result of the sample merchant; Sequentially connect the main business category and the word segmentation results of the sample merchants to obtain first sample information; Sequentially connect the segmentation results of the main business category and the sample merchants to obtain second sample information; The first determination model is obtained according to the first sample information, and the second determination model is obtained according to the second sample information.

6. The merchant information determination method according to claim 5, characterized in that: The obtaining the first determination model according to the first sample information, and the obtaining the second determination model according to the second sample information, comprises: Training the first sample information by using the fasttext algorithm to obtain the first determination model; The second sample information is trained by the fasttext algorithm to obtain the second determination model.

7. The merchant information determination method according to claim 6, characterized in that: The step of inputting the word segmentation result of the target merchant into the first determination model to obtain a second main category probability set corresponding to the target merchant includes: According to the word segmentation result of the target merchant and the first determination model, obtaining a matching probability of each main category corresponding to the word segmentation result of the target merchant; The second main category probability set is obtained according to the matching probability of each main category corresponding to the word segmentation result of the target merchant.

8. The merchant information determination method according to claim 6, characterized in that: The step of inputting the segmentation result of the target merchant into a second determination model to obtain a third main category probability set corresponding to the target merchant includes: According to the segmentation result of the target merchant and the second determination model, obtaining a matching probability of each main category corresponding to the segmentation result of the target merchant; The third main category probability set is obtained according to the matching probability of each main category corresponding to the segmentation result of the target merchant.

9. The merchant information determination method according to claim 1, characterized in that: After performing word segmentation processing on the description information of the target merchant to obtain the word segmentation result of the target merchant, and before inputting the word segmentation result of the target merchant into the first determination model to obtain the second main category probability set corresponding to the target merchant, the method further includes: The noise in the word segmentation result of the target merchant is filtered, wherein the noise includes punctuation marks, numbers and stop words.

10. The merchant information determination method according to claim 1, characterized in that: After performing single-character segmentation processing on the description information of the target merchant to obtain the segmentation result of the target merchant, and before inputting the segmentation result of the target merchant into the second determination model to obtain the third main category probability set corresponding to the target merchant, the method further includes: The noise in the segmentation result of the target merchant is filtered, wherein the noise includes punctuation marks, numbers and stop words.

11. The merchant information determination method according to claim 1, characterized in that: The description information of the target merchant includes: the name of the target merchant, the product name of the target merchant and the user evaluation of the target merchant.

12. A device for determining merchant information, characterized in that: include: A first acquisition module is used to acquire description information of a target merchant; An extraction module, used to extract lexical features contained in the description information of the target merchant, wherein the lexical features include a word segmentation result of the target merchant and a segmentation result of the target merchant; A second acquisition module is used to acquire a first main category probability set corresponding to the target merchant according to the lexical feature, wherein the first main category probability set includes matching probabilities of at least two main categories corresponding to the lexical feature; A main category determination module, configured to determine the main category of the target merchant according to the first main category probability set; The step of acquiring a first main category probability set corresponding to the target merchant according to the vocabulary feature includes: Inputting the word segmentation result of the target merchant into the first determination model to obtain a second main category probability set corresponding to the target merchant, wherein the second main category probability set includes the matching probability of each main category corresponding to the word segmentation result of the target merchant, and the first determination model includes a first mapping relationship between the word segmentation result and the matching probability of the main category; Inputting the segmentation result of the target merchant into the second determination model to obtain a third main category probability set corresponding to the target merchant, wherein the third main category probability set includes the matching probability of each main category corresponding to the segmentation result of the target merchant, and the second determination model includes a second mapping relationship between the segmentation result and the matching probability of the main category; The first main category probability set is obtained according to the second main category probability set and the third main category probability set.

13. An electronic device, characterized in that: include: Memory and processor. Memory stores computer programs. Processor executes programs when it runs: Get the description information of the target merchant; Extracting lexical features contained in the description information of the target merchant, wherein the lexical features include a word segmentation result of the target merchant and a segmentation result of the target merchant; Acquire a first main category probability set corresponding to the target merchant according to the lexical feature, wherein the first main category probability set includes matching probabilities of at least two main categories corresponding to the lexical feature; Determining the main category of the target merchant according to the first main category probability set; The step of acquiring a first main category probability set corresponding to the target merchant according to the vocabulary feature includes: Inputting the word segmentation result of the target merchant into the first determination model to obtain a second main category probability set corresponding to the target merchant, wherein the second main category probability set includes the matching probability of each main category corresponding to the word segmentation result of the target merchant, and the first determination model includes a first mapping relationship between the word segmentation result and the matching probability of the main category; Inputting the segmentation result of the target merchant into the second determination model to obtain a third main category probability set corresponding to the target merchant, wherein the third main category probability set includes the matching probability of each main category corresponding to the segmentation result of the target merchant, and the second determination model includes a second mapping relationship between the segmentation result and the matching probability of the main category; The first main category probability set is obtained according to the second main category probability set and the third main category probability set.

14. The electronic device according to claim 13, characterized in that: The extracting of vocabulary features contained in the description information of the target merchant includes: Performing word segmentation processing on the description information of the target merchant to obtain a word segmentation result of the target merchant; The description information of the target merchant is segmented into single characters to obtain a segmentation result of the target merchant.

15. The electronic device according to claim 13, characterized in that: The obtaining the first main category probability set according to the second main category probability set and the third main category probability set includes: Adding the probabilities of the same type of main categories in the second main category probability set and the third main category probability set to obtain a fourth main category probability set; retaining the probabilities of different types of main categories in the second main category probability set and the third main category probability set to obtain a fifth main category probability set; The fourth main category probability set and the fifth main category probability set are combined to obtain the first main category probability set.

16. The electronic device according to claim 13, characterized in that: The determining the main category of the target merchant according to the first main category probability set includes: Determine at least one main category corresponding to a probability value greater than a preset threshold in the first main category probability set; The at least one main category is used as the main category of the target merchant.

17. The electronic device according to claim 13, characterized in that: Before acquiring the first main category probability set corresponding to the target merchant according to the vocabulary feature, the method further includes: Acquire initial information of sample merchants, wherein the initial information includes description information of the sample merchants and marked main categories, and the number of the sample merchants is at least two; Performing word segmentation processing on the description information of the sample merchant to obtain the word segmentation result of the sample merchant; Performing single-character segmentation processing on the description information of the sample merchant to obtain a segmentation result of the sample merchant; Sequentially connect the main business category and the word segmentation results of the sample merchants to obtain first sample information; Sequentially connect the segmentation results of the main business category and the sample merchants to obtain second sample information; The first determination model is obtained according to the first sample information, and the second determination model is obtained according to the second sample information.

18. The electronic device according to claim 17, characterized in that: The obtaining the first determination model according to the first sample information, and the obtaining the second determination model according to the second sample information, comprises: The second sample information is trained by using the fasttext algorithm to obtain the first determination model; The second sample information is trained by the fasttext algorithm to obtain the second determination model.

19. The electronic device according to claim 18, characterized in that: The step of inputting the word segmentation result of the target merchant into the first determination model to obtain a second main category probability set corresponding to the target merchant includes: According to the word segmentation result of the target merchant and the first determination model, obtaining a matching probability of each main category corresponding to the word segmentation result of the target merchant; The second main category probability set is obtained according to the matching probability of each main category corresponding to the word segmentation result of the target merchant.

20. The electronic device according to claim 18, characterized in that The step of inputting the segmentation result of the target merchant into a second determination model to obtain a third main category probability set corresponding to the target merchant includes: According to the segmentation result of the target merchant and the second determination model, obtaining a matching probability of each main category corresponding to the segmentation result of the target merchant; The third main category probability set is obtained according to the matching probability of each main category corresponding to the segmentation result of the target merchant.

21. The electronic device according to claim 13, characterized in that: After performing word segmentation processing on the description information of the target merchant to obtain the word segmentation result of the target merchant, and before inputting the word segmentation result of the target merchant into the first determination model to obtain the second main category probability set corresponding to the target merchant, the method further includes: The noise in the word segmentation result of the target merchant is filtered, wherein the noise includes punctuation marks, numbers and stop words.

22. The electronic device according to claim 13, characterized in that: After performing single-character segmentation processing on the description information of the target merchant to obtain the segmentation result of the target merchant, and before inputting the segmentation result of the target merchant into the second determination model to obtain the third main category probability set corresponding to the target merchant, the method further includes: The noise in the segmentation result of the target merchant is filtered, wherein the noise includes punctuation marks, numbers and stop words.

23. The electronic device according to claim 13, characterized in that: The description information of the target merchant includes: the name of the target merchant, the product name of the target merchant and the user evaluation of the target merchant.

24. A non-volatile storage medium for storing a computer-readable program, characterized in that: The computer-readable program is used for enabling a computer to execute the merchant information determination method according to any one of claims 1 to 11.