A network security-based evaluation and analysis system

Abstract

The present invention discloses an evaluation and analysis system based on network security, which relates to the technical field of shopping systems. It includes a data collection module, a data preprocessing module, a feature extraction module, a model training module, and an evaluation and analysis module. The feature extraction module includes multiple recognition frames. Through the commodity positioning, feature extraction is performed on the commodities in the picture, and effective information is quickly extracted from all the information of the acquired commodities. The effective information includes commodity brand, commodity price, commodity date, commodity shelf life, and commodity use. Through the text and commodity cutting algorithm, the brands, logos, and texts carried by the commodities in the picture are all disassembled, and feature extraction is performed on the commodity body. Through channel matching, the prices of the commodities identified by the recognition frames in other channels are searched and displayed on the user interface for the user to obtain, enabling the user to purchase a worthy commodity with a calm attitude, avoiding impulse consumption by the user, and helping the user make a more informed choice.

Description

Technical Field

[0001] The present invention relates to the technical field of shopping systems, and particularly to an evaluation and analysis system based on network security. Background Art

[0002] Live commerce is the act of selling goods during a live stream in the live entertainment industry. The emergence of live commerce is due to the popularity of mobile Internet and the fact that young people generally spend a lot of time on their mobile phones every day and are keen on watching live streams. Some early e-commerce companies began to cooperate with live streamers. The live streamers help merchants sell goods and get a commission, which is the origin of live commerce. However, the popularity of the live commerce industry has not raised the threshold of this industry, allowing low-quality goods to be sold through live commerce, resulting in the situation of harming consumers and making the problem of online shopping security more and more serious.

[0003] When selling goods through live stream, the live streamer will set a fixed duration after a product is put on the shelf. Only by purchasing the product within the fixed duration of the live streamer can the consumer enjoy the discount. Within this fixed duration, consumers do not have enough time to compare the prices of the same product in other shopping channels, and may not even have enough time to carefully read the product manual to understand the product, resulting in impulse purchases by consumers without sufficient understanding of the product. Summary of the Invention

[0004] The purpose of the present invention is to provide an evaluation and analysis system based on network security to solve the problems raised in the above background art.

[0005] To solve the above technical problems, the present invention provides the following technical solution: An evaluation and analysis system based on network security, including a data collection module, a data preprocessing module, a feature extraction module, a model training module, and an evaluation and analysis module. The feature extraction module includes multiple recognition frames, and at least one recognition frame is used to recognize pictures. The recognition process of the recognition frame includes a product positioning unit, a text and product cutting algorithm, and channel matching.

[0006] Through the product positioning, feature extraction is performed on the product in the picture, and effective information is quickly extracted from all the information of the product obtained. The effective information includes product brand, product price, product date, product shelf life, and product use.

[0007] Through the text and product cutting algorithm, the brand, logo, and text carried by the product in the picture are all disassembled, and feature extraction is performed on the product body.

[0008] Through channel matching, the prices of the products recognized by the recognition frame in other channels are searched and displayed on the user interface for the user to obtain.

[0009] According to the above technical solution, the data acquisition module includes a preferential information extraction unit, multiple commodity sales channel price information extraction units, a live commodity information extraction unit, and a commodity sales video extraction unit;

[0010] The preferential information extraction unit obtains the preferential calculation method of the live commodity, directly helping users obtain the way to purchase at the lowest price;

[0011] The multiple commodity sales channel price information extraction units obtain the prices of the same commodity in other channels during the same period;

[0012] The live commodity information extraction unit obtains all the information of the live commodity, and all the information includes the extraction of all the text on the commodity introduction page;

[0013] The commodity sales video extraction unit obtains the entire process video of the live sale of the commodity.

[0014] According to the above technical solution, the commodity sales video extraction unit has an automatic deletion function, and the commodity sales video extraction unit selects a segmented video recording method to divide the entire live sales video into several short videos of equal duration through the segmented video recording method.

[0015] According to the above technical solution, the data preprocessing module includes a time node reservation unit;

[0016] The time node reservation unit frames the live time period for users to purchase commodities, divides the entire live time period into a user purchase commodity confirmation time period and a non-confirmation time period, and gives a priority level division for the data preprocessing module to process data by dividing the time period.

[0017] According to the above technical solution, the data preprocessing module further includes a commodity information data cleaning unit and a commodity information data modularization unit;

[0018] The commodity information data cleaning unit cleans the information obtained by the data acquisition module to ensure that the information obtained by the data acquisition module is all available information and avoid data such as garbled characters;

[0019] The commodity information data modularization unit analyzes the information collected by the data acquisition module and obtains the same type of commodities and commodity prices in other channels.

[0020] According to the above technical solution, the model training module includes a learning library, a confirmation unit, and an evaluation unit;

[0021] Record the user's shopping habits through the learning library. The same type of products are grouped into a whole, and the user's shopping habits are extracted from the same type of products. The shopping habits include two categories: price selection and product selection. The price selection habit is determined by the model training module by analyzing whether the user will prioritize price when purchasing the same type of products. The product selection habit is determined by the model training module by analyzing whether the user will prioritize product quality when purchasing the same type of products;

[0022] Update the user's shopping habits through the confirmation unit;

[0023] Divide the user's shopping habit record data stored in the learning library by time period:

[0024] L1, confirm the time period duration;

[0025] L2, learn the user's shopping habits in the previous 3 time periods in the learning library, and analyze and confirm the user's shopping habits;

[0026] L3, analyze the user's shopping habits recorded in the 4th time period, confirm the user's shopping habits, and compare them with the previously confirmed user's shopping habits. If the two user's shopping habits are the same, execute according to the latest confirmed user's shopping habits. If the two user's shopping habits are different, execute according to the old shopping habits. Starting from the 4th time period, record the user's shopping habits in 3 time periods through the learning library again, and re-analyze and confirm the user's shopping habits;

[0027] L4, help the user confirm whether the selected product conforms to the shopping habit by confirming the user's purchase record.

[0028] According to the above technical solution, the evaluation unit includes a test scoring unit and a tracing unit;

[0029] Regularly conduct shopping habit tests on users through the test scoring unit. The shopping habit test is to manually design a questionnaire page for users to test;

[0030] Determine the time point of each shopping habit test through the tracing unit. Each time after recording the user's shopping habits in 3 time periods through the learning library again, conduct a shopping habit test on the user. Double-confirm the accurate positioning of the user's shopping habits through the results obtained from the analysis of the user's shopping habits by the model training module and the results of the user's evaluation. In case of result deviation, the user's evaluation results shall prevail.

[0031] According to the above technical solution, the evaluation and analysis module includes evaluation results and analysis reports;

[0032] The obtained evaluation results are provided for users to quickly understand whether this product needs to be purchased. Some products sold through live broadcasts have timeliness;

[0033] The analysis report is provided for users to understand the advantages and disadvantages of the products being sold.

[0034] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: In the present invention, when users watch live sales, the product information is saved by taking screenshots, and the pictures are quickly scanned and recognized through the recognition frame to obtain all the information of the product. The information such as non-attached products and product prices from different channels is displayed for users to view. Users can quickly browse the information and, combined with the results given by the evaluation and analysis module, choose whether to purchase such products. Through the rational results given by the evaluation and analysis module, the influence brought by the live broadcast environment to users is reduced. By using big data analysis and machine learning technologies, the value of products can be quickly and accurately evaluated, enabling users to purchase a worthy product with a calm attitude, avoiding impulse consumption by users, and helping users make more informed choices. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The drawings are used to provide a further understanding of the present invention and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention and do not constitute a limitation to the present invention. In the drawings:

[0036] Figure 1 is a schematic diagram of the basic process for product evaluation of the present invention;

[0037] Figure 2 is a schematic diagram of the product evaluation and analysis system of the present invention;

[0038] Figure 3 is a schematic diagram of the detailed process for product evaluation of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0039] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0040] Please refer to Figures 1-3 , the present invention provides a technical solution: An evaluation and analysis system based on network security, including a data acquisition module, a data preprocessing module, a feature extraction module, a model training module, and an evaluation and analysis module. The feature extraction module includes multiple recognition frames, and at least one recognition frame is used to recognize pictures. The recognition process of the recognition frame includes a product positioning unit, a text and product cutting algorithm, and channel matching;

[0041] Extract the features of the products in the picture through product positioning, and quickly extract the effective information from all the information of the acquired products. The effective information includes product brand, product price, product date, product shelf life, and product use.

[0042] Disassemble the brand, logo, and text carried by the products in the picture through the text and product cutting algorithm, and extract the features of the product body.

[0043] Search and display the prices of the products identified by the recognition frame in other channels through channel matching on the user interface for the user to obtain.

[0044] For example, for a shirt product, disassemble the brand, logo, and text on the shirt product to get a shirt without attachments. The recognition frame recognizes this shirt without attachments, obtains the price of this shirt without attachments for the user to obtain, and compares the price of the shirt without attachments with that of the shirt product, giving the user multiple choices. The disadvantage of live streaming sales is the low after-sales guarantee, many low-quality products, and the existence of purchasing plain-colored clothes without any additional signs at a low price, and then randomly pricing them through additional signs, making consumers buy low-quality products at a high price. Compare the additional products and the products without attachments through the recognition frame for the user to choose.

[0045] The data acquisition module includes a preferential information extraction unit, multiple product sales channel price information extraction units, a live broadcast product information extraction unit, and a product sales video extraction unit.

[0046] Obtain the preferential calculation method of the live broadcast products through the preferential information extraction unit, and directly help the user obtain the way to purchase at the lowest price.

[0047] Obtain the prices of the same products in other channels during the same period through multiple product sales channel price information extraction units.

[0048] Obtain all the information of the products being live broadcast through the live broadcast product information extraction unit. All the information includes the extraction of all the text on the product introduction page.

[0049] Obtain the entire video of the product sales process of the products being live broadcast through the product sales video extraction unit.

[0050] The product sales video extraction unit has an automatic deletion function. The product sales video extraction unit selects the split video recording method. Through the split video recording method, the entire live streaming sales video is divided into several short videos of equal duration, which is convenient for users to search and delete / keep the videos later.

[0051] The data preprocessing module includes a time node reservation unit.

[0052] The live broadcast time period for the user to purchase goods is framed by the time node reservation unit. The entire live broadcast time period is divided into a user purchase goods confirmation time period and a non-confirmation time period. By dividing the time period, a priority level division for the data preprocessing module to process data is given. The user purchase goods confirmation time period is the video time period for live selling the goods purchased by the user. The retention of this video time period can be used as evidence for after-sales rights protection. The non-confirmation time period is when the user watches the live selling video but has no purchase behavior. This video time period can be used as the object when the data preprocessing module cleans data. The data preprocessing module regularly cleans data to ensure that the database space meets normal use and does not harm the interests of users.

[0053] The data preprocessing module also includes a commodity information data cleaning unit and a commodity information data modularization unit;

[0054] The commodity information data cleaning unit cleans the information obtained by the data acquisition module to ensure that all the information obtained by the data acquisition module is available information and avoid data such as garbled characters;

[0055] The commodity information data modularization unit analyzes the information collected by the data acquisition module and obtains the same type of commodities and commodity prices in other channels.

[0056] The model training module includes a learning library, a confirmation unit, and an evaluation unit;

[0057] The learning library records the habits of users purchasing goods. The same type of goods is summarized as a whole, and the purchasing habits of users are extracted from the same type of goods. The purchasing habits include two categories: price selection and product selection. The price selection habit is determined by the model training module by analyzing whether the user will take price as the first priority selection criterion when purchasing the same type of goods, that is, when purchasing goods multiple times, the probability of selecting low-price goods in the same type of goods exceeds 85%, and the user is determined to be a price-oriented user. The product selection habit is determined by the model training module by analyzing whether the user will take product quality as the first priority selection criterion when purchasing the same type of goods, that is, when purchasing goods multiple times, the probability of selecting high-price goods in the same type of goods exceeds 60%, and the user is determined to be a product-oriented user;

[0058] The confirmation unit updates the purchasing habits of users;

[0059] The recorded data of the purchasing habits of users stored in the learning library is divided according to time periods:

[0060] L1, the duration of the confirmation time period;

[0061] L2, the purchasing habits of users in the previous 3 time periods recorded in the learning library are analyzed to confirm the purchasing habits of users;

[0062] L3. Analyze the recorded user purchase habits during the fourth time period to confirm the user's purchase habits. Compare them with the previously confirmed user purchase habits. If the two user purchase habits are the same, execute according to the newly confirmed user purchase habits. If there are differences between the two user purchase habits, execute according to the old purchase habits. Starting from the fourth time period, record the user purchase habits in the learning library for another three time periods and re-analyze and confirm the user's purchase habits;

[0063] L4. By confirming the user's purchase records, help the user confirm whether the selected product conforms to the purchase habits. Products that conform to the user's purchase habits have a lower probability of impulse consumption and are more in line with the user's purchase habits.

[0064] The evaluation unit includes a test scoring unit and a tracing unit;

[0065] Regularly conduct purchase habit tests on users through the test scoring unit. The purchase habit test is to manually design a questionnaire page for users to test;

[0066] Determine the time point of each purchase habit test through the tracing unit. After each time of recording the user's purchase habits in the learning library for three time periods, conduct a purchase habit test on the user. Through the model training module, double-confirm the results obtained from the analysis of the user's purchase habits and the results of the user evaluation to accurately locate the user's purchase habits. In case of result deviation, take the user evaluation result as the standard.

[0067] The evaluation and analysis module includes evaluation results and analysis reports;

[0068] Through the obtained evaluation results, enable the user to quickly understand whether this product needs to be purchased. Some products sold through live broadcasts have timeliness. By displaying the evaluation results, the user can quickly know the answer and avoid impulse consumption caused by the user's consideration within a short period of time;

[0069] Through the analysis report, enable the user to understand the advantages and disadvantages of the products being sold. For products without timeliness, the user can first understand the advantages and disadvantages of the products and then make a choice;

[0070] Through the final evaluation and analysis module, reduce the probability of user impulse consumption and at the same time solve the problem of selling inferior products to users through the form of live selling.

[0071] When users watch live sales, they save product information by taking screenshots, quickly scan and recognize the pictures through the recognition frame to obtain all the information of the products, display the products without additional items, prices of products from different channels, etc. for users to view. Users quickly browse the information and, combined with the results given by the evaluation and analysis module, choose whether to purchase such products. Through the rational results given by the evaluation and analysis module, the impact brought by the live broadcast environment to users can be reduced. By using big data analysis and machine learning technologies, the value of products can be evaluated quickly and accurately, enabling users to purchase a worthy product with a calm attitude, avoiding impulse consumption by users, and helping users make more informed choices.

[0072] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device.

[0073] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A network security-based evaluation and analysis system, comprising a data acquisition module, a data preprocessing module, a feature extraction module, a model training module and an evaluation and analysis module, characterized in that: The feature extraction module includes a plurality of recognition frames, at least one of which is used to recognize the image, and the recognition process of the recognition frame includes a product positioning unit, a text and product segmentation algorithm, and a channel matching; Extract features of the product in the image through product positioning, and quickly extract valid information from all the acquired product information, wherein the valid information includes product brand, product price, product date, product shelf life, and product use; The brand, logo and text of the product in the picture are separated by the text and product segmentation algorithm, and the product features are extracted. Through channel matching, the price search of the product identified by the identification box in other channels is displayed on the user interface for users to obtain; The data collection module includes a preferential information extraction unit, a plurality of commodity sales channel price information extraction units, a live commodity information extraction unit, and a commodity sales video extraction unit; The preferential information extraction unit obtains the preferential calculation method of the live broadcast product, directly helping the user to obtain the method of how to purchase at the lowest price; Acquire the price of the same commodity in other channels within the same time period through the multiple commodity selling channel price information extraction units; The live broadcast product information extraction unit is used to obtain all information about the product being broadcast live, including all text extraction of the product introduction page; The commodity selling video extraction unit obtains the live broadcast of the whole process of commodity selling; The commodity selling video extraction unit has an automatic deletion function. The commodity selling video extraction unit selects to adopt a separated video recording method, and divides the entire live selling video into several short videos of equal length through the separated video recording method; The data preprocessing module includes a time node reservation unit; The live broadcast time period for users to purchase goods is framed by the time node reservation unit, and the entire live broadcast time period is divided into a time period for users to confirm the purchase of goods and a time period without confirmation. By dividing the time periods, the data preprocessing module is given priority for processing data; The model training module includes a learning library, a confirmation unit and an evaluation unit; The learning library records the user's habit of purchasing goods, and the same type of goods are grouped as a whole. The user's purchasing habits are extracted from the same type of goods. The purchasing habits include price selection and product selection. The price selection habit is determined by the model training module by analyzing whether the user will take price as the first priority when purchasing the same type of goods. The product selection habit is determined by the model training module by analyzing whether the user will take product quality as the first priority when purchasing the same type of goods. Updating the user's purchasing habits through the confirmation unit; Divide the user purchase habit record data stored in the learning library into time periods: L1, confirm the time period duration; L2, the learning library records the user's purchasing habits in the previous three time periods and analyzes and confirms the user's purchasing habits; L3: Analyze the user purchase habits recorded in the fourth time period, confirm the user's purchase habits, and compare them with the previously confirmed user purchase habits. If the two users' purchase habits are consistent, execute according to the latest confirmed user purchase habits. If the two users' purchase habits are different, execute according to the old purchase habits. Take the fourth time period as the starting time period, record the user purchase habits in the three time periods through the learning library again, and re-analyze and confirm the user's purchase habits. L4, by confirming the user's purchase history, helps the user confirm whether the selected product is in line with their purchase habits; The evaluation and analysis module includes evaluation results and analysis reports; The evaluation results provide users with a quick understanding of whether they need to buy the product. Some products sold through live streaming are time-sensitive. The analysis report allows users to understand the advantages and disadvantages of the products being sold.

2. The network security evaluation and analysis system according to claim 1 is characterized in that: The data preprocessing module also includes a commodity information data clearing unit and a commodity information data modularization unit; The information obtained by the data collection module is cleaned by the commodity information data cleaning unit to ensure that the information obtained by the data collection module is usable information and to avoid the appearance of garbled data; The commodity information data modular unit analyzes the information collected by the data collection module, and obtains commodities and commodity prices of the same type in other channels.

3. The network security evaluation and analysis system according to claim 2 is characterized in that: The evaluation unit includes a test scoring unit and a tracing unit; The test scoring unit is used to regularly test the user's purchasing habits. The purchasing habit test is to manually design a questionnaire page for the user to test. The tracing unit is used to determine the time point of each purchase habit test. After the user's purchase habits within three time periods are recorded again through the learning library each time, a purchase habit test is conducted on the user. The model training module is used to double-confirm the results of the user's purchase habit analysis and the results of the user evaluation to accurately locate the user's purchase habits. In case of result deviation, the user evaluation result shall prevail.

Images