Intelligent bidding method for scrap steel transaction

Abstract

The invention discloses an intelligent bidding method for scrap steel transaction, and belongs to the technical field of scrap steel transaction bidding. The problem that an existing steel scrap transaction bidding platform is single in bidding mode is solved, optimization and processing are carried out on matching logic, original single matching is evolved into multivariate matching logic, a label algorithm, a subscription algorithm, an attention algorithm and a spacing algorithm are introduced between a bonding user and the platform, and the bidding mode of the steel scrap transaction bidding platform is optimized and processed. And performing weighted average calculation on the result of the calculation logic to form a comprehensive algorithm, thereby improving the accuracy of resource pushing, improving the perceptibility of the user, forming user resonance and improving the value of the resource.

Description

Technical Field

[0001] The present invention relates to the technical field of scrap steel trading bidding, and in particular to an intelligent bidding method for scrap steel trading. Background Art

[0002] Scrap steel refers to scrap steel produced in industrial production or daily life, which can be recycled after processing and sorting. Scrap steel trading bidding refers to the process of determining the price and transaction of scrap steel through bidding. However, the existing scrap steel trading bidding has the following problems:

[0003] 1. In scrap steel transactions, traditional bidding methods often rely on manual operations, and there are problems such as information asymmetry and opaque bidding. The pain point of inconsistent quality inspection and penalty standards has brought certain challenges to scrap-producing and scrap-using enterprises in the transaction process.

[0004] 2. The platform provides equal business opportunities for both buyers and sellers through online transactions by introducing competitive bidding. It also provides a massive market, which ensures that waste-producing enterprises can sell at high prices and solves the cost-effectiveness of procurement for waste-using enterprises. However, the platform's bidding model is single, users lack labeling operations, the resource transaction rate is not high, the business value-added is not obvious, business opportunities are seriously lost, and the bid evaluation is not transparent enough.

[0005] 3. Although the existing intelligent bidding system can improve bidding efficiency, it lacks comprehensive consideration of multiple factors, resulting in inaccurate bidding results that still require manual intervention.

[0006] 4. The rights and interests of both buyers and sellers cannot be guaranteed. For businesses transacted online, the platform has weak control and dispute resolution capabilities, and is unable to move beyond the acquaintance model, resulting in bottlenecks in business promotion.

[0007] 5. The platform lacks labeled operations for users, and relevant information cannot be delivered to users in a timely and effective manner, resulting in a certain lag in business.

[0008] Therefore, the existing demand is not met, and we propose an intelligent bidding method for scrap steel trading. Summary of the invention

[0009] The purpose of the present invention is to provide an intelligent bidding method for scrap steel trading. By optimizing and processing the matching logic, the original single matching is evolved into a multi-matching logic. A label algorithm, a subscription algorithm, a follow-up algorithm, and a spacing algorithm are introduced between the bonding users and the platform. The results of the above calculation logics are weighted averaged to form a comprehensive algorithm, thereby improving the accuracy of resource push and solving the problems raised in the above background technology.

[0010] To achieve the above object, the present invention provides the following technical solution: a scrap steel trading intelligent bidding method, comprising the following steps:

[0011] Data collection: collect user data and business data. User data refers to the information of platform users, including basic information, transaction information and behavior information. Business data refers to the bidding data released by the platform, including business attributes, session attributes and rule attributes. The collected user data and business data are stored in the data processing center;

[0012] Among them, basic information includes geographic location coordinates, name, mobile phone number and ID card;

[0013] Transaction information, including transaction type, transaction quantity, transaction amount and transaction frequency;

[0014] Behavioral information, including page views, operation paths, subscriptions, and followers;

[0015] Business attributes, including seller, variety, specification and source location;

[0016] Session attributes, including session name, contact person, contact information, and deposit;

[0017] Rule attributes, including bidding mode, bidding time, bidding cycle and bidding rules;

[0018] Data feature extraction: Extract features from user data and business data stored in the data processing center to extract features that affect bidding;

[0019] Algorithm model calculation: Based on the extracted features, an intelligent bidding algorithm model is constructed to match the distance to the source of goods and user preferences. The distance between the resource and the user is calculated through the distance matching of the source of goods, and the business data that meets the user characteristics is intelligently sorted through user preference matching;

[0020] Parameter optimization configuration: Provide parameter setting services according to business needs, including user configuration, business configuration and basic configuration;

[0021] Model algorithm execution: After completing the above steps, the user successfully publishes the bidding session and initiates the intelligent bidding method, which comprehensively considers distance, preference, attention and subscription factors to push the best sessions to users so that they can respond, evaluate costs, profits and business cycles, and facilitate transactions in a timely manner.

[0022] Furthermore, user data is collected through cooperation with the platform, and business data is collected through API interfaces.

[0023] Furthermore, the data processing center provides the following functions:

[0024] Data query and analysis function: Provides real-time data query and analysis functions, including data screening, statistical analysis and report generation, so that users can quickly obtain and analyze user data and business data stored in the data processing center;

[0025] Data export and sharing function: Provide data export and sharing functions, including data format conversion, export report and sharing link, so that users can export and share user data and business data stored in the data processing center;

[0026] Data assurance function: Provides security mechanisms for data, including data encryption, access control, and data backup and recovery, to ensure the security and reliability of user data and business data stored in the data processing center.

[0027] Furthermore, the data protection function is specifically:

[0028] Data encryption: user data and business data are encrypted and stored through symmetric or asymmetric encryption methods, and users decrypt the data using keys;

[0029] Access control: Control the user's access rights to user data and business data by verifying the user's identity and role, and prevent unauthorized access and theft by providing different operation permissions to different user roles;

[0030] Data backup and recovery: regularly copy user data and business data to a safe location to prevent the original data from being lost or damaged. If user data and business data are lost or damaged, restore the user data and business data from the backup data.

[0031] Furthermore, data feature extraction is specifically as follows:

[0032] The data processing center will classify the collected user data, extract user operation preferences, locations, operation paths and browsing frequencies, and label and group users;

[0033] The collected business data will be classified and processed at the data processing center, and the varieties, rules and deposit data in the business data will be extracted. The bidding sessions will be labeled and classified, and grouped according to varieties, sources of goods and deposits.

[0034] Furthermore, the algorithm model calculation is specifically as follows:

[0035] Distance matching of the supply source: using the user's current location coordinates and the supply source coordinates, calculate the distance between the resource and the user, and perform intelligent sorting in the data processing center;

[0036] User preference matching: intelligently sort the business data that meets the user's characteristics based on the user's preferences, distance, attention, and subscription weights.

[0037] Furthermore, the intelligent bidding algorithm model constructed by the algorithm model calculation includes a collaborative filtering algorithm, wherein the collaborative filtering algorithm includes user-based collaborative filtering, item-based collaborative filtering and content-based recommendation algorithm, specifically:

[0038] User-based collaborative filtering: recommend the content that the user likes to the target user by finding other users with similar interests to the target user. For user u and user v, the set of items they have paid attention to together is I, and user u’s attention to item i is r. ui , user v’s attention to variety i is r vi , the average attention of user u and user v are r ui and r vi , then the Pearson correlation coefficient between user u and user v is calculated as:

[0039]

[0040] Where: u and v represent different users;

[0041] i represents a set of varieties that do not need to be paid attention to;

[0042] r ui represents the attention of user u to variety i;

[0043] r vi represents the attention of v users to variety i;

[0044] Item-based collaborative filtering: Send the content of the show to the target user in the form of recommendation. The similarity between items is calculated using methods such as Pearson correlation coefficient or cosine similarity. Taking cosine similarity as an example, for item i and item j, the set of users who have paid attention to them together is U ij , user u’s attention to item i is r i , the attention paid to item j is r j , variety i and variety attention vector are r i and r j , then the cosine similarity calculation formula between variety i and variety j is:

[0045]

[0046] In the formula: i and j represent different items;

[0047] r i Represents the user's attention to item i;

[0048] r j Represents the user's attention to variety j;

[0049] The content-based recommendation algorithm matches the characteristics of the session data with the user's historical behavior and recommends content that matches the user's interests. The specific formula is as follows:

[0050]

[0051] In the formula: i represents the category;

[0052] j represents the place of origin;

[0053] a generation attention;

[0054] r ai Represents the attention paid to the category.

[0055] Furthermore, the parameter optimization configuration is as follows:

[0056] User configuration: According to user preferences and characteristics, the frequency, time, efficiency and cycle of matching bidding session push are set for users to ensure rapid information delivery;

[0057] Business configuration: Provide targeted services for bidding sessions, set sales scope and tag users according to business characteristics and transaction rules to ensure business dissemination;

[0058] Basic configuration: Provides basic service configuration, flexible and variable label configuration, and forms an intelligent algorithm.

[0059] Compared with the prior art, the present invention has the following beneficial effects:

[0060] The present invention optimizes and processes the matching logic, evolving the original single matching into a multi-matching logic, and introducing a tag algorithm, a subscription algorithm, a follow-up algorithm, and a spacing algorithm to bond users and platforms. The results of the above calculation logics are weighted averaged to form a comprehensive algorithm, thereby improving the accuracy of resource push, enhancing user perception, forming user resonance, and increasing the value of resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 This is a system flow chart of the intelligent bidding method for scrap steel trading of the present invention;

[0062] Figure 2 This is a business relationship diagram of the intelligent bidding method for scrap steel trading of the present invention. DETAILED DESCRIPTION

[0063] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0064] In order to solve the problems of information asymmetry and opaque bidding in the existing bidding methods, the single bidding model of the platform, the lack of comprehensive consideration of multiple factors, resulting in inaccurate bidding results, and the weak platform control and dispute handling capabilities for online transactions, the inability to get out of the acquaintance model, the bottleneck of business promotion, the lack of labeling operation of platform users, and the inability to effectively convey relevant information to users in a timely manner, please refer to Figure 1-Figure 2 , this embodiment provides the following technical solutions:

[0065] A scrap steel trading intelligent bidding method comprises the following steps:

[0066] Data collection: Collect user data and business data, that is, collect information such as the buyer's business scope, the seller's customer group, the current bid, the location distance between the delivery place and the seller's area, etc. User data refers to the information of platform users, including basic information, transaction information and behavior information. Business data refers to the bidding data released by the platform, including business attributes, session attributes and rule attributes. The collected user data and business data are stored in the data processing center;

[0067] It should be noted that user data is collected through cooperation with the platform. The collection of user data requires informing the user and obtaining the user's consent. Business data is collected through API interfaces.

[0068] Among them, basic information includes geographic location coordinates, name, mobile phone number and ID card;

[0069] Transaction information, including transaction type, transaction quantity, transaction amount and transaction frequency;

[0070] Behavioral information, including page views, operation paths, subscriptions, and followers;

[0071] Business attributes, including seller, variety, specification and source location;

[0072] Session attributes, including session name, contact person, contact information, and deposit;

[0073] Rule attributes, including bidding mode, bidding time, bidding cycle and bidding rules;

[0074] Data feature extraction: Extract features from user data and business data stored in the data processing center to extract features that affect bidding;

[0075] Algorithm model calculation: Based on the extracted features, an intelligent bidding algorithm model is constructed to match the distance to the source of goods and user preferences. The distance between the resource and the user is calculated through the distance matching of the source of goods, and the business data that meets the user characteristics is intelligently sorted through user preference matching;

[0076] Parameter optimization configuration: Provide parameter setting services according to business needs, including user configuration, business configuration and basic configuration; configure model parameters according to actual needs, and optimize and adjust them; provide optimal resource push according to the set index algorithm to improve resource screening efficiency; use margin services to protect the transaction rights and interests of both buyers and sellers; use online quotations and intelligent bid evaluation to avoid integrity risks; use online services to solve the pain point of price opacity in the scrap steel industry;

[0077] Model algorithm execution: After completing the above steps, the user successfully publishes the bidding session and initiates the intelligent bidding method, which comprehensively considers distance, preference, attention and subscription factors to push the best sessions to users so that they can respond, evaluate costs, profits and business cycles, and facilitate transactions in a timely manner.

[0078] The technical effect of the above content is: by collecting user data and business data, performing feature extraction and algorithm model calculation, to achieve efficient and accurate bidding matching. The feature extraction of user data and business data involves basic information such as geographic location, name, mobile phone number, ID number, etc., transaction information such as transaction type, transaction quantity, transaction amount, transaction frequency, and behavioral information such as access path, subscription and attention. After in-depth analysis and processing, this information can be used to mine user preferences and behavior patterns. In the algorithm model calculation stage, by extracting and matching the features of user data and business data, the distance matching of the source of goods and the user preference matching are achieved. This matching can accurately reflect the user's purchasing intention and purchasing habits, and help improve the efficiency and quality of the transaction. In addition, the present invention also provides a parameter setting service, which can perform user configuration, business configuration and basic configuration according to different business needs. These configurations can be automatically optimized by intelligent algorithms to adapt to the ever-changing business environment. In the model algorithm execution stage, after the user successfully publishes the bidding session, the intelligent bidding method will be started. By comprehensively considering factors such as distance, preference, attention and subscription, the best quality sessions are pushed to users so that users can respond in time, evaluate costs, profits and business cycles, and thus promote the completion of transactions.

[0079] The data processing center provides the following functions:

[0080] Data query and analysis function: Provides real-time data query and analysis functions, including data screening, statistical analysis and report generation, so that users can quickly obtain and analyze user data and business data stored in the data processing center;

[0081] Data export and sharing function: Provide data export and sharing functions, including data format conversion, export report and sharing link, so that users can export and share user data and business data stored in the data processing center;

[0082] Data assurance function: Provides security mechanisms for data, including data encryption, access control, and data backup and recovery, to ensure the security and reliability of user data and business data stored in the data processing center.

[0083] The technical effect of the above content is: through the data query and analysis function, users can query and analyze the user data and business data stored in the data processing center in real time, and users can quickly locate and view data through various screening conditions (such as time, category, region, etc.), and can also perform statistical analysis and report generation, so that users can deeply understand and use these data. The data export and sharing function allows users to export user data and business data in the data processing center to local or cloud files or databases, and can also share data with others. When exporting data, the data processing center will convert the data format to ensure that the exported data can be used normally on different devices. In addition, the export report function can generate PDF, Excel The data processing center can generate reports in different formats, which is convenient for users to print, share or further analyze the data. The data protection function is to ensure the security and reliability of user data and business data stored in the data processing center. The data processing center has adopted a variety of protection measures. First, the data is encrypted, and only users with corresponding permissions can decrypt and use the data. Secondly, strict control is implemented on data access. Finally, the data processing center also regularly backs up data to prevent data loss. If data leakage or damage occurs, the data processing center can quickly restore data from the backup. The data processing center is an intelligent tool that integrates data query, analysis, export and sharing. It can help users manage and use data more effectively and enhance the ability of data-driven decision-making.

[0084] Data protection functions, specifically:

[0085] Data encryption: user data and business data are encrypted and stored through symmetric or asymmetric encryption methods, and users decrypt the data using keys;

[0086] Access control: Control the user's access rights to user data and business data by verifying the user's identity and role, and prevent unauthorized access and theft by providing different operation permissions to different user roles;

[0087] Data backup and recovery: regularly copy user data and business data to a safe location to prevent the original data from being lost or damaged. If user data and business data are lost or damaged, restore the user data and business data from the backup data.

[0088] The technical effects of the above content are as follows: the data processing center uses symmetric encryption or asymmetric encryption methods to encrypt and store user data and business data, and the user decrypts the data through the key. This method can effectively prevent the data from being stolen or tampered with without authorization. The data processing center verifies the user's identity and role, controls the user's access rights to user data and business data, and prevents unauthorized access and theft by providing different operation permissions for different user roles. For example, ordinary users may only have the functions of viewing and exporting data, while administrators can perform more operations, such as adding data, modifying data, etc. The data processing center regularly copies user data and business data to a safe location to prevent the original data from being lost or damaged. When user data and business data are lost or damaged, the user data and business data can be restored from the backup data, thereby avoiding the loss of original data due to system failure, human error, etc. By taking a variety of security measures, the security and reliability of user data and business data stored in the data processing center are effectively guaranteed.

[0089] Data feature extraction, specifically:

[0090] The data processing center will classify the collected user data, extract user operation preferences, locations, operation paths and browsing frequencies, and label and group users;

[0091] The collected business data will be classified and processed at the data processing center, and the varieties, rules and deposit data in the business data will be extracted. The bidding sessions will be labeled and classified, and grouped according to varieties, sources of goods and deposits.

[0092] The technical effects of the above content are: classifying user data, extracting user operation preferences, locations, operation paths and browsing frequencies, and labeling and grouping users, which can explore users' interests and behavior patterns, so as to more accurately meet users' personalized needs in the subsequent recommendation process, classifying business data, extracting varieties, rules and margin data from business data, and labeling and classifying bidding sessions. Grouping by variety, source of goods and margin can understand the attributes of various bidding sessions, so as to better match user preferences and needs when recommending. By analyzing user behavior data and business data, finding out the rules and patterns, and visualizing and labeling them, the data can be made more structured and easy to understand. The data feature extraction results can not only improve the accuracy of subsequent recommendation systems, but also help the operation team better understand users, optimize and adjust recommendation strategies.

[0093] The algorithm model calculation is as follows:

[0094] Distance matching of the supply source: using the user's current location coordinates and the supply source coordinates, calculate the distance between the resource and the user, and perform intelligent sorting in the data processing center;

[0095] User preference matching: intelligently sort the business data that meets the user's characteristics based on the user's preferences, distance, attention, and subscription weights.

[0096] The technical effect of the above content is: in the algorithm model calculation link, the user's current location coordinates and the source coordinates are first used to calculate the distance between the resource and the user, and then they are intelligently sorted, which can ensure that the products that meet the user's needs are pushed to the user in the shortest time. Then, according to the user's preferences, distance, attention and subscription weight ratio, the business data that meets the user's characteristics is intelligently sorted, so that the products and services that they are most interested in can be recommended based on their needs and preferences.

[0097] The intelligent bidding algorithm model constructed by algorithm model calculation includes a collaborative filtering algorithm, wherein the collaborative filtering algorithm includes user-based collaborative filtering, item-based collaborative filtering and content-based recommendation algorithm, specifically:

[0098] User-based collaborative filtering: recommend the content that the user likes to the target user by finding other users with similar interests to the target user. For user u and user v, the set of items they have paid attention to together is I, and user u’s attention to item i is r. ui , user v’s attention to variety i is r vi , the average attention of user u and user v are r ui and r vi, then the Pearson correlation coefficient between user u and user v is calculated as:

[0099]

[0100] Where: u and v represent different users;

[0101] i represents a set of varieties that do not need to be paid attention to;

[0102] r ui represents the attention of user u to variety i;

[0103] r vi represents the attention of v users to variety i;

[0104] Item-based collaborative filtering: Send the content of the session to the target user in the form of recommendation. The similarity between items is calculated using methods such as Pearson correlation coefficient or cosine similarity. Taking cosine similarity as an example, for item i and item j, the set of users who have paid attention to them together is U ij , user u’s attention to item i is r i , the attention paid to item j is r j , variety i and variety attention vector are r i and r j , then the cosine similarity calculation formula between variety i and variety j is:

[0105]

[0106] In the formula: i and j represent different items;

[0107] r i Represents the user's attention to item i;

[0108] r j Represents the user's attention to variety j;

[0109] The content-based recommendation algorithm matches the characteristics of the session data with the user's historical behavior and recommends content that matches the user's interests. The specific formula is as follows:

[0110]

[0111] In the formula: i represents the category;

[0112] j represents the place of origin;

[0113] a generation attention;

[0114] r ai Represents the attention paid to the category.

[0115] The technical effect of the above content is: through optimization and processing on the matching logic, the original single matching has evolved into a multi-matching logic, and the label algorithm, subscription algorithm, attention algorithm, and spacing algorithm have been introduced to bond users and platforms. The results of the above calculation logic are weighted averaged to form a comprehensive algorithm, thereby improving the accuracy of resource push.

[0116] As shown in Table 1 and Table 2, the data structure design is divided into a feature table (Table 1) and a push table (Table 2). The feature table mainly stores user trading preferences, product labels, and behavior characteristics, and the push table mainly stores push objects, push types, and push content.

[0117] Table 1

[0118]

[0119]

[0120] Table 2

[0121]

[0122]

[0123]

[0124] Parameter optimization configuration, specifically:

[0125] User configuration: According to user preferences and characteristics, the frequency, time, efficiency and cycle of matching bidding session push are set for users to ensure rapid information delivery;

[0126] Business configuration: Provide targeted services for bidding sessions, set sales scope and tag users according to business characteristics and transaction rules to ensure business dissemination;

[0127] Basic configuration: Provides basic service configuration, flexible and variable label configuration, and forms an intelligent algorithm.

[0128] The technical effect of the above content is: in the parameter optimization configuration stage, the frequency, time, efficiency and cycle of the matching bidding session push will be set for the user according to the user's behavior pattern, interests, hobbies, value orientation and other information to ensure the rapid reach of information, which is the so-called "personalized recommendation". For business configuration, targeted services will be provided for bidding sessions based on specific business characteristics and transaction rules, such as setting specific promotion strategies, advertising delivery rules, price strategies, etc. At the same time, sales scope and label users will be set for bidding sessions to more accurately target customers and promote. In terms of basic configuration, a series of basic service configurations will be provided, including user management, product management, scene management, etc., to ensure the smooth progress of the entire bidding process. At the same time, flexible and variable label configurations will also be provided so that users and merchants can adjust and optimize their bidding strategies at any time according to actual conditions. All of the above configurations will be integrated into the intelligent algorithm model, which will run automatically during each bidding process to adjust and optimize push strategies, bidding strategies, etc. in real time to achieve the best bidding effect.

[0129] Working principle: By optimizing and processing the matching logic, the original single matching has evolved into a multi-matching logic. The tag algorithm, subscription algorithm, attention algorithm and spacing algorithm are introduced to bond users and platforms. The results of the above calculation logic are weighted averaged to form a comprehensive algorithm, thereby improving the accuracy of resource push, enhancing user perception, forming user resonance and increasing the value of resources.

[0130] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0131] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. A scrap steel trading intelligent bidding method, characterized in that: The following steps are involved: Data collection: collect user data and business data. User data refers to the information of platform users, including basic information, transaction information and behavior information. Business data refers to the bidding data released by the platform, including business attributes, session attributes and rule attributes. The collected user data and business data are stored in the data processing center; Among them, basic information includes geographic location coordinates, name, mobile phone number and ID card; Transaction information, including transaction type, transaction quantity, transaction amount and transaction frequency; Behavioral information, including page views, operation paths, subscriptions, and followers; Business attributes, including seller, variety, specification and source location; Session attributes, including session name, contact person, contact information, and deposit; Rule attributes, including bidding mode, bidding time, bidding cycle and bidding rules; Data feature extraction: Extract features from user data and business data stored in the data processing center to extract features that affect bidding; Algorithm model calculation: Based on the extracted features, an intelligent bidding algorithm model is constructed to match the distance to the source of goods and user preferences. The distance between the resource and the user is calculated through the distance matching of the source of goods, and the business data that meets the user characteristics is intelligently sorted through user preference matching; Parameter optimization configuration: Provide parameter setting services according to business needs, including user configuration, business configuration and basic configuration; Model algorithm execution: After completing the above steps, the user successfully publishes the bidding session and initiates the intelligent bidding method, which comprehensively considers distance, preference, attention and subscription factors to push the best sessions to users so that they can respond, evaluate costs, profits and business cycles, and facilitate transactions in a timely manner.

2. The intelligent bidding method for scrap steel trading according to claim 1, characterized in that: User data is collected through cooperation with the platform, and business data is collected through API interfaces.

3. The intelligent bidding method for scrap steel trading according to claim 1 is characterized in that: The data processing center provides the following functions: Data query and analysis function: Provides real-time data query and analysis functions, including data screening, statistical analysis and report generation, so that users can quickly obtain and analyze user data and business data stored in the data processing center; Data export and sharing function: Provide data export and sharing functions, including data format conversion, export report and sharing link, so that users can export and share user data and business data stored in the data processing center; Data assurance function: Provides security mechanisms for data, including data encryption, access control, and data backup and recovery, to ensure the security and reliability of user data and business data stored in the data processing center.

4. The intelligent bidding method for scrap steel trading according to claim 3 is characterized by: The data protection function is specifically: Data encryption: user data and business data are encrypted and stored through symmetric or asymmetric encryption methods, and users decrypt the data using keys; Access control: Control the user's access rights to user data and business data by verifying the user's identity and role, and prevent unauthorized access and theft by providing different operation permissions to different user roles; Data backup and recovery: regularly copy user data and business data to a safe location to prevent the original data from being lost or damaged. If user data and business data are lost or damaged, restore the user data and business data from the backup data.

5. The intelligent bidding method for scrap steel trading according to claim 1 is characterized by: The data feature extraction is specifically as follows: The data processing center will classify the collected user data, extract user operation preferences, locations, operation paths and browsing frequencies, and label and group users; The collected business data will be classified and processed at the data processing center, and the varieties, rules and deposit data in the business data will be extracted. The bidding sessions will be labeled and classified, and grouped according to varieties, sources of goods and deposits.

6. The intelligent bidding method for scrap steel trading according to claim 1 is characterized by: The algorithm model calculation is specifically as follows: Distance matching of the supply source: using the user's current location coordinates and the supply source coordinates, calculate the distance between the resource and the user, and perform intelligent sorting in the data processing center; User preference matching: intelligently sort the business data that meets the user's characteristics based on the user's preferences, distance, attention, and subscription weights.

7. The intelligent bidding method for scrap steel trading according to claim 1, characterized in that: The intelligent bidding algorithm model constructed by the algorithm model calculation includes a collaborative filtering algorithm, wherein the collaborative filtering algorithm includes user-based collaborative filtering, item-based collaborative filtering and content-based recommendation algorithm, specifically: User-based collaborative filtering: recommend the content that the user likes to the target user by finding other users with similar interests to the target user. For user u and user v, the set of items they have paid attention to together is I, and user u’s attention to item i is r. ui , user v’s attention to variety i is r vi , the average attention of user u and user v are r ui and r vi , then the Pearson correlation coefficient between user u and user v is calculated as: Where: u and v represent different users; i represents a set of varieties that do not need to be paid attention to; r ui represents the attention of user u to variety i; r vi represents the attention of v users to variety i; Item-based collaborative filtering: Send the content of the session to the target user in the form of recommendation. The similarity between items is calculated using methods such as Pearson correlation coefficient or cosine similarity. Taking cosine similarity as an example, for item i and item j, the set of users who have paid attention to them together is U ij , user u’s attention to item i is r i , the attention paid to item j is r j , variety i and variety attention vector are r i and r j , then the cosine similarity calculation formula between variety i and variety j is: In the formula: i and j represent different items; r i Represents the user's attention to item i; r j Represents the user's attention to variety j; The content-based recommendation algorithm matches the characteristics of the session data with the user's historical behavior and recommends content that matches the user's interests. The specific formula is as follows: In the formula: i represents the category; j represents the place of origin; a generation attention; r ai Represents the attention paid to the category.

8. The intelligent bidding method for scrap steel trading according to claim 1 is characterized by: The parameter optimization configuration is specifically as follows: User configuration: According to user preferences and characteristics, the frequency, time, efficiency and cycle of matching bidding session push are set for users to ensure rapid information delivery; Business configuration: Provide targeted services for bidding sessions, set sales scope and tag users according to business characteristics and transaction rules to ensure business dissemination; Basic configuration: Provides basic service configuration, flexible and variable label configuration, and forms an intelligent algorithm.

Images