CN110599195A - Method for identifying bill swiping - Google Patents
Method for identifying bill swiping Download PDFInfo
- Publication number
- CN110599195A CN110599195A CN201910713859.XA CN201910713859A CN110599195A CN 110599195 A CN110599195 A CN 110599195A CN 201910713859 A CN201910713859 A CN 201910713859A CN 110599195 A CN110599195 A CN 110599195A
- Authority
- CN
- China
- Prior art keywords
- comment
- comments
- word
- written
- neural network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/335—Filtering based on additional data, e.g. user or group profiles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/018—Certifying business or products
- G06Q30/0185—Product, service or business identity fraud
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Business, Economics & Management (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Entrepreneurship & Innovation (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Finance (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to a method for identifying a bill, which comprises the following steps: acquiring a large number of effective comment text sets of a large number of users, and training word vectors of each word of each comment text; combining two comment texts and the identification value of whether the comment texts belong to the same user at random for multiple times to form a training set; building and training a convolutional neural network for predicting whether the two comment texts are written by the same person; if two comment texts of a large number of different users are predicted to be commented by the same person, a merchant is predicted to have a behavior of swiping. The invention utilizes the characteristics that a person speaks and types and has certain individual language and format style habits, utilizes the thought to further identify the bill swiping, further identifies whether the comment text is written by the same person or not by learning the language habits carried in the comment text through detail characteristics by means of a neural network model, and finally identifies whether a certain online shop has the bill swiping behavior or not through a certain method based on the model.
Description
Technical Field
The invention relates to the field of machine learning, in particular to a method for recognizing a swipe bill.
Background
The technology is widely applied to daily life along with the rapid development of computer technology, wherein online shopping is particularly convenient, electronic shopping is the mainstream shopping mode of the modern times, and in the case that no real object is seen, how to find a good object? is currently, most consumers can see comments left by the purchased users, and a certain reference value is provided for the consumers, but for the merchants, aiming at the common psychology of the buyers, the merchants can take certain measures to improve the business amount of the merchants, the conventional method is to swipe a single sheet, the sheet swiping is quite common, for example, the merchants hiring a plurality of real users to swipe a sheet, the professional swiping a single person to swipe a plurality of account numbers, and the like are all the performances of the consumers, how to recognize the sheet swiping behaviors appears to be necessary.
Disclosure of Invention
In order to solve the defects that in the prior art, a professional person who brushes a bill uses a plurality of account numbers to brush the bill generally exists and is more and more difficult to find, the invention provides a method for identifying the bill.
In order to solve the technical problems, the technical scheme of the invention is as follows:
a method of identifying a swipe ticket, comprising the steps of:
step S1: obtaining effective comment texts of users in the e-commerce website to form an effective comment text set;
step S2: training each effective comment text to obtain a word vector of each word in the effective comment text;
step S3: randomly screening out two comments for many times in the effective comment text set, and combining the two comments screened out each time to form a training sample;
step S4: building a one-dimensional convolutional neural network, wherein an input layer of the one-dimensional convolutional neural network is a word vector of each word in two comment texts in a training sample, and an output layer of the one-dimensional convolutional neural network is a possibility prediction value of whether the two comment texts in the training sample are written by the same person or not, and training the one-dimensional convolutional neural network to obtain an optimal model;
step S5: inputting a new training sample into the optimal model, and predicting whether two comments in the training sample are written by the same person;
step S6: selecting an E-commerce shop, obtaining effective comment texts of all transactions of the E-commerce shop, pairing the effective comment texts of different users pairwise, and respectively inputting the paired effective comment texts into an optimal model for prediction to obtain a prediction result;
step S7: and analyzing the prediction result and judging whether the electric shop has a brushing single row.
Preferably, the step of obtaining the effective comment text of the user in the website of the e-commerce is as follows:
screening a plurality of users, then screening their shopping comments, filtering out all invalid comments, wherein the invalid comments comprise: the system automatically reviews, commonly used word reviews and empty reviews, and retains the remaining user review text.
Preferably, in the step S2, for each piece of valid comment text, no filtering word is made,
Punctuation and special format processing, all texts are reserved, the length of each effective comment text is increased to the sample length of the maximum word number in the data set, blank positions are filled with specified characters, the input dimensions of samples are consistent, and word vectors of each word are trained.
Preferably, in step S4, the specific process of building the one-dimensional convolutional neural network is as follows:
the input layer of the one-dimensional convolutional neural network is two comment texts with the same length represented by word vectors, the number of channels is the size of the word vectors, the activation function of the output layer is a Sigmoid activation function, the cost function adopts a classical Cross entry function, and the output layer is a possibility prediction value for whether the two comment texts are written by the same person or not.
Preferably, in step S5, if the two comments belong to the same user, the identification value is set to 1, which indicates that the two comments are written by the same person; if the two comments are from different users, the identification value is set to 0, indicating that the two comments were not written by the same person.
Preferably, in step S6, a threshold value g is set, and if it is predicted that the ratio of the number of combinations written by the same person to all the combinations is greater than the set threshold value g, it is determined that the electric shop is laid with the list.
Preferably, the threshold value g is set to 0.3.
Compared with the prior art, the technical scheme of the invention has the beneficial effects that:
the method utilizes the characteristic that the style habits of speaking and typing of each person are difficult to change, and the language style habits of each comment are difficult to be different because a large number of comments are needed when the same person swipes the comments for many times, so that the model can identify whether the comments are written by the same person through details and characteristics, and further identify whether a certain online shop has the behavior of swiping the comments. In addition, when the convolutional neural network is used for learning, the feature extraction operation can be completely handed to a machine, and the one-dimensional data can be regarded as two-dimensional data with the number of lines being 1, so that the convolutional neural network used by the method can automatically extract the features of the comment language, and the features can better help the convolutional neural network to learn the language style and habit.
Drawings
FIG. 1 is a flow chart of a method of the present invention;
FIG. 2 is a flow chart of constructing a training set according to the present invention;
FIG. 3 is a flow chart of how the invention identifies a brush order using a trained model;
FIG. 4 is a schematic diagram of a one-dimensional convolutional neural network structure according to the present invention.
Detailed Description
The drawings are for illustrative purposes only and are not to be construed as limiting the patent;
for the purpose of better illustrating the embodiments, certain features of the drawings may be omitted, enlarged or reduced, and do not represent the size of an actual product;
it will be understood by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted.
The technical solution of the present invention is further described below with reference to the accompanying drawings and examples.
Example 1
As shown in fig. 1 to 4, a method for recognizing a swipe sheet includes the following steps:
step S1: obtaining effective comment texts of users in the e-commerce website to form an effective comment text set;
step S2: training each effective comment text to obtain a word vector of each word in the effective comment text;
step S3: randomly screening out two comments for many times in the effective comment text set, and combining the two comments screened out each time to form a training sample;
step S4: building a one-dimensional convolutional neural network, wherein an input layer of the one-dimensional convolutional neural network is a word vector of each word in two comment texts in a training sample, and an output layer of the one-dimensional convolutional neural network is a possibility prediction value of whether the two comment texts in the training sample are written by the same person or not, and training the one-dimensional convolutional neural network to obtain an optimal model;
step S5: inputting a new training sample into the optimal model, and predicting whether two comments in the training sample are written by the same person;
step S6: selecting an E-commerce shop, obtaining effective comment texts of all transactions of the E-commerce shop, pairing the effective comment texts of different users pairwise, and respectively inputting the paired effective comment texts into an optimal model for prediction to obtain a prediction result;
step S7: and analyzing the prediction result and judging whether the electric shop has a brushing single row.
As a preferred embodiment, the step of obtaining the effective comment text of the user in the website of the commercial website is as follows:
screening a plurality of users, then screening their shopping comments, filtering out all invalid comments, wherein the invalid comments comprise: the system automatically reviews, commonly used word reviews and empty reviews, and retains the remaining user review text.
As a preferred embodiment, as shown in fig. 2, in step S2, for each valid comment text, no processing of filtering words, punctuation marks, and special formats is performed, the text is all retained, the length of each valid comment text is increased to the sample length of the maximum number of words in the data set, the blank positions are filled with designated characters, the input dimensions of the samples are made uniform, and a word vector is trained for each word.
As a preferred embodiment, in step S4, a specific process of building a one-dimensional convolutional neural network is as follows:
as shown in fig. 4, the input layer of the one-dimensional convolutional neural network is two comment texts with the same length represented by word vectors, the number of channels is the size of the word vectors, the output layer activation function is a Sigmoid activation function, the cost function adopts a classical Cross entry function, and the output layer is a possibility prediction value of whether the two comment texts are written by the same person.
As a preferred embodiment, in step S5, if the two comments belong to the same user, the identification value is set to 1, which indicates that the two comments are written by the same person; if the two comments are from different users, the identification value is set to 0, indicating that the two comments were not written by the same person.
As a preferred embodiment, in step S6, a threshold g is set, and if it is predicted that the ratio of the number of combinations written by the same person to all the combinations is greater than the set threshold g, it is determined that the electric shop is laid with the brushing list.
As a preferred embodiment, the threshold value g is set to 0.3.
Example 2
The invention relates to a method for identifying a bill, which is implemented by the general flow chart shown in figure 1 and comprises the following steps:
step S1: obtaining effective comment texts of users in the e-commerce website to form an effective comment text set;
step S2: training each effective comment text to obtain a word vector of each word in the effective comment text;
step S3: randomly screening out two comments for many times in the effective comment text set, and combining the two comments screened out each time to form a training sample;
step S4: building a one-dimensional convolutional neural network, wherein an input layer of the one-dimensional convolutional neural network is a word vector of each word in two comment texts in a training sample, and an output layer of the one-dimensional convolutional neural network is a possibility prediction value of whether the two comment texts in the training sample are written by the same person or not, and training the one-dimensional convolutional neural network to obtain an optimal model;
step S5: inputting a new training sample into the optimal model, and predicting whether two comments in the training sample are written by the same person;
step S6: selecting an E-commerce shop, obtaining effective comment texts of all transactions of the E-commerce shop, pairing the effective comment texts of different users pairwise, and respectively inputting the paired effective comment texts into an optimal model for prediction to obtain a prediction result;
step S7: and analyzing the prediction result and judging whether the electric shop has a brushing single row.
As a preferred embodiment, in step 1, a set of valid comment texts of a large number of users in the website of the commercial is obtained, that is, a plurality of users are screened, then, their shopping comments are screened, invalid comments such as all system automatic comments, commonly used comments, empty comments and the like are filtered, and the remaining comment texts of the users are retained and further used as a training set.
As a preferred embodiment, in step 2, for each valid comment, no word filtering, punctuation, special formatting processing is performed, the text is all retained, the length of each comment sample is increased to the sample length with the maximum number of words in the data set, the blank position is filled with specified characters, the input dimensions of the samples are made consistent, and a word vector is trained for each word.
As a preferred embodiment, in step 3, a large number of valid comment samples with consistent lengths, which are represented by word vectors, are obtained, and two comment text pairs are randomly selected multiple times, that is, two comment samples are combined into a new training sample. If the two comment samples belong to the same user, setting the possibility prediction value as 1 to indicate that the two comment texts are written by the same person; if the two comment samples are from different users, then the likelihood predictor is set to 0, indicating that they are not written by the same person.
5 samples were generated as a demonstration example, as shown in Table 1, A, B, C being users, the numbers representing their comments
| Combination of | Whether it is written by the same person |
| (comment A1, comment B1) | 0 |
| (A review 1, A review 2) | 1 |
| (B review 1, B review 5) | 1 |
| (A comment 1, C comment 2) | 0 |
| (C comment 3, B comment 2) | 0 |
TABLE 1
The relevant flow of constructing the training set is shown in fig. 2.
As a preferred embodiment, in the step 4, a structural schematic diagram of a one-dimensional convolutional neural network is constructed, as shown in fig. 4, an input layer is two comment samples with the same length represented by word vectors, the number of channels is the size of the word vectors, if the above is set, the number of words of each comment sample is unified to 200, the dimension of the word vector of each word is 16, since the input layer includes two comment texts, and the dimension of the input layer is 400 × 16, it should be noted that, among them, the first 200 × 16 node belongs to a first comment text, and the last 200 × 16 node belongs to a second comment text, that is, if the words of the first comment text do not occupy 200, the specified characters are required to be filled instead of filling the second comment text immediately, and the second comment text can only start from 201; the output layer activation function is a Sigmoid activation function, the cost function can adopt a classic cross Encopy function, the training process of the network is not described in detail, the most common back propagation algorithm can be used for feeding back the tuning parameters, and the hyper-parameters can be determined by a Grid Search method of a Scikt-Learn framework.
The output layer activation function is a Sigmoid function, and if the output value exceeds a set threshold value, two comment texts are predicted to be written by the same person; otherwise the prediction is not.
As a preferred embodiment, in step 6, to identify whether there is a list-swiping behavior of an e-shop, the relevant flowchart is shown in fig. 3, and the identification process is as follows: the method comprises the steps of firstly obtaining effective comments of all users obtained by a shop, detecting that the same user swipes a single line to the same merchant for many times because the same user may buy the comments for many times in the shop, so that the significance is not provided, only combining comment texts of different users, pairing the comment texts in pairs, and using the combined comment texts as a data set to be predicted, inputting a group of the comment texts into a model to predict whether two comment texts of each group are written by the same person or not. A threshold value g (e.g. 0.3) is set, and if the ratio of the number of combinations written by the same person to all the number of combinations is predicted to be greater than the set threshold value g, the shop owner can be predicted to have a row for brushing the list.
The same or similar reference numerals correspond to the same or similar parts;
the terms describing positional relationships in the drawings are for illustrative purposes only and are not to be construed as limiting the patent;
it should be understood that the above-described embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. Other variations and modifications will be apparent to persons skilled in the art in light of the above description. And are neither required nor exhaustive of all embodiments. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the claims of the present invention.
Claims (7)
1. A method for identifying a swipe ticket, comprising the steps of:
step S1: obtaining effective comment texts of users in the e-commerce website to form an effective comment text set;
step S2: training each effective comment text to obtain a word vector of each word in the effective comment text;
step S3: randomly screening out two comments for many times in the effective comment text set, and combining the two comments screened out each time to form a training sample;
step S4: building a one-dimensional convolutional neural network, wherein an input layer of the one-dimensional convolutional neural network is a word vector of each word in two comment texts in a training sample, and an output layer of the one-dimensional convolutional neural network is a possibility prediction value of whether the two comment texts in the training sample are written by the same person or not, and training the one-dimensional convolutional neural network to obtain an optimal model;
step S5: inputting a new training sample into the optimal model, and predicting whether two comments in the training sample are written by the same person;
step S6: selecting an E-commerce shop, obtaining effective comment texts of all transactions of the E-commerce shop, pairing the effective comment texts of different users pairwise, and respectively inputting the paired effective comment texts into an optimal model for prediction to obtain a prediction result;
step S7: and analyzing the prediction result and judging whether the electric shop has a brushing single row.
2. The method for identifying the billing of claim 1, wherein in the step S1, the step of obtaining the effective comment text of the user in the website of the electronic commerce is as follows:
screening a plurality of users, then screening their shopping comments, filtering out all invalid comments, wherein the invalid comments comprise: the system automatically reviews, commonly used word reviews and empty reviews, and retains the remaining user review text.
3. The method for recognizing a swipe according to claim 1, wherein in step S2, for each valid comment text, no processing for filtering words, punctuation marks, and special formats is performed, the text is all kept, the length of each valid comment text is increased to a sample length of the maximum number of words in the data set, blank positions are filled with designated characters, the input dimensions of the samples are made consistent, and a word vector is trained for each word.
4. The method for identifying the billing of claim 1, wherein in the step S4, the specific process of constructing the one-dimensional convolutional neural network is as follows:
the input layer of the one-dimensional convolutional neural network is two comment texts with the same length represented by word vectors, the number of channels is the size of the word vectors, the activation function of the output layer is a Sigmoid activation function, the cost function adopts a classical crossEncopy function, and the output layer is a possibility prediction value for whether the two comment texts are written by the same person or not.
5. The method for recognizing the swipe note of claim 1, wherein in step S5, if the two comments belong to the same user, the identification value is set to 1, which indicates that the two comments are written by the same person; if the two comments are from different users, the identification value is set to 0, indicating that the two comments were not written by the same person.
6. The method for identifying a billing according to claim 1, wherein in step S6, a threshold g is set, and if it is predicted that the ratio of the number of combinations written by the same person to all the combinations is greater than the set threshold g, the billing is determined to be present in the e-shop.
7. The method of claim 6, wherein the threshold g is set to 0.3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910713859.XA CN110599195B (en) | 2019-08-02 | 2019-08-02 | Method for identifying bill swiping |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910713859.XA CN110599195B (en) | 2019-08-02 | 2019-08-02 | Method for identifying bill swiping |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110599195A true CN110599195A (en) | 2019-12-20 |
| CN110599195B CN110599195B (en) | 2023-04-14 |
Family
ID=68853421
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910713859.XA Active CN110599195B (en) | 2019-08-02 | 2019-08-02 | Method for identifying bill swiping |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110599195B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116484841A (en) * | 2023-05-10 | 2023-07-25 | 道有道科技集团股份公司 | Information verification system and method based on automatic auditing |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109902179A (en) * | 2019-03-04 | 2019-06-18 | 上海宝尊电子商务有限公司 | The method of screening electric business comment spam based on natural language processing |
-
2019
- 2019-08-02 CN CN201910713859.XA patent/CN110599195B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109902179A (en) * | 2019-03-04 | 2019-06-18 | 上海宝尊电子商务有限公司 | The method of screening electric business comment spam based on natural language processing |
Non-Patent Citations (4)
| Title |
|---|
| 李璐？等: "虚假评论检测研究综述", 《计算机学报》 * |
| 李静: "基于卷积神经网络的虚假评论的识别", 《软件》 * |
| 皮琪等: "基于深度学习的虚假评论识别", 《网络新媒体技术》 * |
| 韩俣谈: "基于卷积神经网络的虚假评论识别模型", 《科学技术创新》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116484841A (en) * | 2023-05-10 | 2023-07-25 | 道有道科技集团股份公司 | Information verification system and method based on automatic auditing |
| CN116484841B (en) * | 2023-05-10 | 2023-10-03 | 道有道科技集团股份公司 | Information verification system and method based on automatic auditing |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110599195B (en) | 2023-04-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107977798B (en) | Risk assessment method for quality of electronic commerce product | |
| CN103729359B (en) | A kind of method and system recommending search word | |
| US11586658B2 (en) | Method and device for matching semantic text data with a tag, and computer-readable storage medium having stored instructions | |
| CN108388660B (en) | Improved E-commerce product pain point analysis method | |
| CN106803168A (en) | A kind of abnormal transfer accounts method for detecting and device | |
| CN105183717B (en) | A kind of OSN user feeling analysis methods based on random forest and customer relationship | |
| CN104778186B (en) | Merchandise items are mounted to the method and system of standardized product unit | |
| KR102218468B1 (en) | Answer recommendation system and method based on text content and emotion analysis | |
| CN108845986A (en) | A kind of sentiment analysis method, equipment and system, computer readable storage medium | |
| CN107886231B (en) | Service quality evaluation method and system for customer service | |
| CN107833059B (en) | Service quality evaluation method and system for customer service | |
| TW201905736A (en) | Information push method and system | |
| CN109582788A (en) | Comment spam training, recognition methods, device, equipment and readable storage medium storing program for executing | |
| CN109815480B (en) | Data processing method and device and storage medium | |
| CN110858219A (en) | Logistics object information processing method and device and computer system | |
| CN112434501A (en) | Work order intelligent generation method and device, electronic equipment and medium | |
| CN116401379A (en) | Financial product data pushing method, device, equipment and storage medium | |
| Gandhe et al. | Sentiment analysis of Twitter data with hybrid learning for recommender applications | |
| CN104077288B (en) | Web page contents recommend method and web page contents recommendation apparatus | |
| CN110599195B (en) | Method for identifying bill swiping | |
| CN110968670B (en) | Method, device, equipment and storage medium for acquiring attributes of popular commodities | |
| CN114969498A (en) | Method and device for recommending industrial faucet information | |
| CN112884480A (en) | Method and device for constructing abnormal transaction identification model, computer equipment and medium | |
| CN111666410B (en) | Emotion classification method and system for commodity user comment text | |
| CN108763258A (en) | Document subject matter parameter extracting method, Products Show method, equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |