Sodium hydrosulfite price trend prediction method and system

Abstract

The invention relates to the field of data processing, in particular to a sodium hydrosulfite price trend prediction method and system. The method comprises the following steps: acquiring multi-dimensional transaction data of sodium hydrosulfite; establishing a real-time transaction network model of sodium hydrosulfite based on the multi-dimensional transaction data; the real-time transaction network model is provided with a plurality of networking modes, and each networking mode is constructed by transaction data of a corresponding dimension; receiving a query instruction of a target object to obtain attribute information of the target object; analyzing the real-time monitoring data flow corresponding to the attribute information in the real-time transaction network model by adopting a change trend prediction model to obtain sodium hydrosulfite transaction trend change characteristics matched with the target object; and generating and displaying corresponding sodium hydrosulfite price change trend information to the target object based on the transaction trend change characteristics. According to the method, sodium hydrosulfite price trend prediction can be realized under the conditions of data missing and poor accuracy, and the reliability and timeliness of price prediction are improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular, to a method and system for predicting the price trend of sodium dithionite. Background Art

[0002] Sodium dithionite, commonly known as Rongalite, is a white sand-like crystal or light yellow powder. It has strong reducibility and is extremely easy to be oxidized in the air. It will become ineffective or even may burn when it is affected by moisture or exposed to the air. When heated to 190 °C, it will decompose and release sulfur dioxide and a large amount of heat, with the risk of explosion. Sodium dithionite is easily soluble in water and insoluble in ethanol. It is mainly used for pulp bleaching in the paper industry, reduction dyeing and fabric bleaching in the textile printing and dyeing industry, bleaching and anti-corrosion in the food industry, and also has important applications in the pharmaceutical industry, electronic industry, and wastewater treatment, etc. However, when using and storing it, the operating procedures must be strictly followed, and attention should be paid to safety and sealed storage.

[0003] In the related technologies, sodium dithionite is applicable to various fields, and thus, the fluidity of sodium dithionite is relatively strong. For some small and medium-sized enterprises, it is difficult to comprehensively obtain the transaction data of sodium dithionite among each supply and demand party, resulting in a lack of complete data support for price prediction. At the same time, some regional market data may be missing or inaccurate, affecting the judgment of the overall market price trend. For example, the transaction systems among small and medium-sized enterprises usually do not have standardized data interfaces and cannot obtain transaction data by docking with official data interfaces like large enterprises or regulated markets. Even if some enterprises use management software, the data formats and standards among different software are not unified, making it difficult to achieve data sharing and integration.

[0004] Therefore, there is an urgent need to propose a brand-new technical solution to solve at least one of the above technical problems. Summary of the Invention

[0005] In view of the technical problems existing in the prior art, the present invention provides a method and system for predicting the price trend of sodium dithionite, so as to realize the prediction of the price trend of sodium dithionite under the condition of data shortage and poor accuracy, and improve the reliability and timeliness of the price prediction of sodium dithionite.

[0006] In a first aspect, an embodiment of the present application provides a method for predicting the price trend of sodium dithionite, and the method includes:

[0007] Obtain multi-dimensional transaction data of sodium dithionite; the multi-dimensional transaction data at least includes: transaction data in different geographical regions, transaction data of enterprises of different scales, and transaction data on different trading platforms;

[0008] Build a real-time trading network model of sodium hydrosulfite based on the multi-dimensional trading data; the real-time trading network model is set with multiple networking modes, and each networking mode is constructed by using the trading data of the corresponding dimension; the multiple networking modes at least include: geographical area trading network, enterprise trading network, platform trading network, and comprehensive trading network; the comprehensive trading network is formed by networking with trading data of at least two dimensions;

[0009] Receive a query instruction of a target object to obtain the attribute information of the target object; the attribute information at least includes one of the following: the geographical location where the target object is located, the user type to which the target object belongs, the institution to which the target object belongs, and the trading platform available to the target object;

[0010] Use a change trend prediction model to analyze the real-time monitoring data stream corresponding to the attribute information in the real-time trading network model to obtain the change characteristics of the sodium hydrosulfite trading trend matched by the target object;

[0011] Generate and display the corresponding sodium hydrosulfite price change trend information to the target object based on the trading trend change characteristics.

[0012] In a second aspect, an embodiment of the present application provides a sodium hydrosulfite price trend prediction system, which includes the following units, where,

[0013] An acquisition unit configured to obtain multi-dimensional trading data of sodium hydrosulfite; the multi-dimensional trading data at least includes: trading data of different geographical areas, trading data of enterprises of different scales, and trading data of different trading platforms;

[0014] A networking unit configured to build a real-time trading network model of sodium hydrosulfite based on the multi-dimensional trading data; the real-time trading network model is set with multiple networking modes, and each networking mode is constructed by using the trading data of the corresponding dimension; the multiple networking modes at least include: geographical area trading network, enterprise trading network, platform trading network, and comprehensive trading network; the comprehensive trading network is formed by networking with trading data of at least two dimensions;

[0015] An interaction unit configured to receive a query instruction of a target object to obtain the attribute information of the target object; the attribute information at least includes one of the following: the geographical location where the target object is located, the user type to which the target object belongs, the institution to which the target object belongs, and the trading platform available to the target object;

[0016] A prediction unit configured to use a change trend prediction model to analyze the real-time monitoring data stream corresponding to the attribute information in the real-time trading network model to obtain the change characteristics of the sodium hydrosulfite trading trend matched by the target object;

[0017] The interaction unit is further configured to generate and display corresponding sodium dithionite price change trend information to the target object based on the transaction trend change characteristics.

[0018] In a third aspect, an embodiment of the present application provides an electronic device, which includes:

[0019] At least one processor, a memory, and an input-output unit;

[0020] Wherein, the memory is used to store a computer program, and the processor is used to call the computer program stored in the memory to execute the sodium dithionite price trend prediction method in the first aspect.

[0021] In a fourth aspect, a computer-readable storage medium is provided, which includes instructions that, when the instructions are run on a computer, cause the computer to execute the sodium dithionite price trend prediction method in the first aspect.

[0022] The beneficial effects of the present invention are as follows: A sodium dithionite price trend prediction method and system are provided. In this technical solution, first, multi-dimensional transaction data of sodium dithionite is obtained; the multi-dimensional transaction data at least includes: transaction data in different geographical regions, transaction data of enterprises of different scales, and transaction data of different trading platforms. Furthermore, a real-time transaction network model of sodium dithionite is established based on the multi-dimensional transaction data; the real-time transaction network model is provided with multiple networking modes, and each networking mode is constructed by using transaction data of the corresponding dimension; the multiple networking modes at least include: geographical region transaction network, enterprise transaction network, platform transaction network, and comprehensive transaction network; the comprehensive transaction network is obtained by networking with transaction data of at least two dimensions. Then, a query instruction of the target object is received to obtain the attribute information of the target object; the attribute information at least includes one of the following: the geographical location where the target object is located, the user type to which the target object belongs, the institution to which the target object belongs, and the trading platform available to the target object. Then, a change trend prediction model is used to analyze the real-time monitoring data stream corresponding to the attribute information in the real-time transaction network model to obtain the sodium dithionite transaction trend change characteristics matched by the target object. Finally, corresponding sodium dithionite price change trend information is generated and displayed to the target object based on the transaction trend change characteristics.

[0023] This technical solution integrates multi-dimensional data from different geographical regions, enterprises of various scales, and trading platforms, presents the overall market picture from multiple perspectives, comprehensively captures price influencing factors, and significantly improves the prediction accuracy. Moreover, based on attribute information such as the geographical location and user type of the target object, personalized price trend information is provided to meet the specific needs of each entity and assist them in making favorable decisions. Furthermore, by establishing a real-time trading network model and analyzing the real-time monitored data stream, the market dynamics are tracked in a timely manner, enabling price predictions to reflect market changes in real time and facilitating participants to adjust their strategies promptly. Finally, the change trend prediction model can also be used to analyze the real-time trading network model to explore the potential relationships between data in different dimensions and discover price influencing factors and patterns that are difficult to detect by traditional methods, such as specific price change rules among different elements in the comprehensive trading network. This technical solution can achieve the price trend prediction of sodium dithionite in the case of data shortage and poor accuracy, and improve the reliability and timeliness of the sodium dithionite price prediction. Brief Description of the Drawings

[0024] Figure 1 is a schematic flowchart of a method and system for predicting the price trend of sodium dithionite according to an embodiment of the present application;

[0025] Figure 2 is a schematic structural diagram of a system for predicting the price trend of sodium dithionite according to an embodiment of the present application;

[0026] Figure 3 is a schematic structural diagram of an electronic device according to an embodiment of the present application;

[0027] Figure 4 is a schematic structural diagram of a medium device according to an embodiment of the present application. Detailed Embodiments

[0028] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present application.

[0029] In the description of the present application, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present application, "a plurality" means two or more unless otherwise specifically defined.

[0030] In the description of the present application, the term "for example" is used to mean "serving as an example, illustration, or explanation". Any embodiment described as "for example" in the present application is not necessarily to be construed as more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to implement and use the present invention. In the following description, details are set forth for purposes of explanation. It should be understood that those of ordinary skill in the art can recognize that the present invention can be implemented without the use of these specific details. In other instances, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but rather to be in line with the broadest scope consistent with the principles and features disclosed in the present application.

[0031] Sodium dithionite, commonly known as sodium hydrosulfite, is a white sand-like crystal or light yellow powder. It has strong reducibility and is extremely easy to be oxidized in the air. It will become ineffective or even may burn when exposed to moisture or in the air, and will decompose and release sulfur dioxide and a large amount of heat when heated to 190 °C, with the risk of explosion. Sodium dithionite is easily soluble in water and insoluble in ethanol. It is mainly used for pulp bleaching in the paper industry, reduction dyeing and fabric bleaching in the textile printing and dyeing industry, bleaching and preservation in the food industry, and also has important applications in the pharmaceutical industry, electronic industry, and wastewater treatment, etc. However, strict operating procedures need to be followed during its use and storage, and attention should be paid to safety and sealed storage.

[0032] In the related art, sodium dithionite is widely used and has strong fluidity. However, there are many technical problems in current price prediction. It is difficult to comprehensively obtain the transaction data of sodium dithionite among small and medium-sized enterprises. Its trading system lacks a standardized data interface. Even if management software is used, due to the inconsistent formats and standards, data sharing and integration are extremely difficult, which results in the lack of complete data support for price prediction. At the same time, there are missing or inaccurate regional market data, which seriously affects the judgment of the overall market price trend. In addition, there is a lag in market information transmission. Price prediction relies on timely and accurate information, such as raw material purchase prices, product sales prices, inventory levels, etc. However, there will be delays in actually obtaining this information, making the prediction results unable to reflect the latest changes in the market in a timely manner. Due to incomplete and inaccurate data, key features cannot be accurately extracted. For example, without important transaction or regional data, it is difficult to identify features such as the periodicity and seasonality of price fluctuations, resulting in incomplete input variables for the data analysis model, which greatly affects the accuracy of the model. Therefore, there is an urgent need to propose a new technical solution to solve problems such as difficult data acquisition, information lag, and inaccurate models, so as to improve the reliability and timeliness of sodium dithionite price prediction.

[0033] In addition, in the related art, there is a certain lag in the transmission of market information of sodium dithionite. Price forecasting often requires timely and accurate market information, such as raw material procurement prices, product sales prices, inventory levels, etc. However, in actual operation, obtaining this information may be delayed, making the forecasting results unable to reflect the latest changes in the market in a timely manner.

[0034] In summary, incomplete and inaccurate data may lead to the inability to accurately extract the key features of the data, affecting the analysis of market price trends. For example, the lack of certain important transaction data or regional data may prevent the accurate identification of characteristics such as the periodicity and seasonality of price fluctuations, making the input variables of the data analysis model incomplete and thus affecting the accuracy of the model.

[0035] Therefore, there is an urgent need to propose a brand-new technical solution to solve at least one of the above technical problems.

[0036] The embodiments of the present application provide a method and system for predicting the price trend of sodium dithionite. In this technical solution, first, multi-dimensional transaction data of sodium dithionite is obtained; the multi-dimensional transaction data at least includes: transaction data of different geographical regions, transaction data of enterprises of different scales, and transaction data of different trading platforms. Furthermore, a real-time transaction network model of sodium dithionite is established based on the multi-dimensional transaction data; the real-time transaction network model is provided with multiple networking modes, and each networking mode is constructed by using the transaction data of the corresponding dimension; the multiple networking modes at least include: geographical region transaction network, enterprise transaction network, platform transaction network, and comprehensive transaction network; the comprehensive transaction network is formed by networking the transaction data of at least two dimensions. Then, a query instruction of a target object is received to obtain the attribute information of the target object; the attribute information at least includes one of the following: the geographical location where the target object is located, the user type to which the target object belongs, the institution to which the target object belongs, and the trading platform available to the target object. Then, a change trend prediction model is used to analyze the real-time monitoring data stream corresponding to the attribute information in the real-time transaction network model to obtain the change characteristics of the sodium dithionite transaction trend matched by the target object. Finally, based on the change characteristics of the transaction trend, the corresponding sodium dithionite price change trend information is generated and displayed to the target object.

[0037] In the embodiments of the present application, on the one hand, by integrating the multi-dimensional transaction data of different geographical regions, different scales of enterprises, and different trading platforms, the overall picture of the sodium dithionite market can be reflected from multiple perspectives, so as to more comprehensively capture the factors affecting prices and significantly improve the accuracy of price trend prediction. For example, the transaction data of geographical regions can reflect the impact of supply and demand differences in different regions on prices. Different enterprise scales may lead to different procurement or sales strategies, which in turn affect prices. Combining these factors can more accurately predict price trends.

[0038] In a second aspect, personalized price trend information is provided based on the attribute information of the target object, such as the geographical location, the type of user it belongs to, the institution it belongs to, and the available trading platforms. Different target objects have different roles and needs in the market, and this personalized service can meet the specific needs of each entity and help them make decisions that are more in line with their own interests. For example, manufacturing enterprises may be more concerned about the price trends in the regions where raw materials are supplied, while traders may be more concerned about the price differences on different trading platforms.

[0039] In a third aspect, the establishment of a real-time trading network model and the analysis of real-time monitoring data streams can timely track the dynamic changes in the market, enabling price trend prediction to reflect the latest market conditions in real time. The market situation changes rapidly, and timely information feedback helps market participants adjust their strategies in a timely manner to cope with market changes.

[0040] In a fourth aspect, by using a change trend prediction model to analyze the real-time trading network model, the potential relationships between data in different dimensions can be explored, and factors and patterns affecting prices that are difficult to detect by traditional methods can be discovered. For example, by analyzing the complex relationships between geographical regions, enterprises, and platforms in the comprehensive trading network, it may be found that when enterprises in certain regions conduct transactions through specific platforms, there are specific price change patterns.

[0041] Furthermore, it should be noted that compared with the technical problem of difficult data acquisition in related technologies, the technical solution of this application obtains multi-dimensional trading data from different geographical regions, enterprises of different scales, and different trading platforms, expanding the scope of data sources. Compared with traditional methods, it covers a wider range of market entities and trading scenarios, effectively solving the problem of incomplete data. For example, data from different geographical regions can make up for some data missing due to regional differences, data from enterprises of different scales can reflect the behaviors of different market participants, and data from different trading platforms cover diverse trading channels, making the data more comprehensive.

[0042] Regarding the technical problem of information lag in related technologies, the technical solution of this application constructs a real-time trading network model with multiple networking modes based on multi-dimensional trading data. In particular, the comprehensive trading network uses data networking in at least two dimensions. This way, data in different dimensions complement and verify each other, helping to timely discover and correct inaccurate data. For example, when there are contradictions between geographical region trading data and enterprise trading data in some aspects, possible incorrect data can be found through comparative analysis and corrected. At the same time, the fusion of data in different dimensions can also provide richer information, further improving the accuracy of the data.

[0043] Here, the real-time trading network model combines with the real-time monitoring data stream to obtain the latest market information in a timely manner. Once inaccurate or incomplete data is found, it can be quickly updated and supplemented. For example, if a large sodium hydrosulfite production enterprise newly emerges in a certain region, the real-time monitoring system can promptly capture this information and incorporate it into the data model, thus ensuring that the data always reflects the latest and accurate situation of the market.

[0044] Regarding the technical problem of model accuracy in related technologies, the technical solution of this application can also screen out the corresponding real-time monitoring data stream from the massive multi-dimensional trading data according to the attribute information of the target object for analysis. This targeted data processing method reduces the interference of irrelevant or incorrect data on the prediction result and improves the accuracy and effectiveness of the data. For example, for a target object that only conducts transactions on a specific trading platform, only focus on the trading data related to this platform to avoid the influence of inaccurate data from other platforms.

[0045] All in all, the technical solution of this application can realize the price trend prediction of sodium hydrosulfite in the case of data loss and poor accuracy, and improve the reliability and timeliness of the sodium hydrosulfite price prediction.

[0046] The sodium hydrosulfite price trend prediction solution provided by the embodiments of this application can also be executed by an electronic device, and this electronic device can be a server, a server cluster, or a cloud server. This electronic device can also be a terminal device such as a mobile phone, a computer, a tablet computer, a wearable device, or a dedicated device (such as a dedicated terminal device with a sodium hydrosulfite price trend prediction system, etc.). These electronic devices can also be equipped with the chips introduced in the above embodiments. Or, these electronic devices can also install a service program for executing the sodium hydrosulfite price trend prediction solution.

[0047] Figure 1 For the flow schematic diagram of a sodium hydrosulfite price trend prediction method and system provided by the embodiments of this application, as Figure 1 shown, the method includes the following steps:

[0048] 101. Obtain the multi-dimensional trading data of sodium hydrosulfite;

[0049] 102. Establish a real-time trading network model of sodium hydrosulfite based on the multi-dimensional trading data;

[0050] 103. Receive a query instruction of the target object to obtain the attribute information of the target object;

[0051] 104. Use a change trend prediction model to analyze the real-time monitoring data stream corresponding to the attribute information in the real-time trading network model to obtain the change characteristics of the sodium hydrosulfite trading trend matching the target object;

[0052] 105. Generate and display the corresponding sodium dithionite price change trend information to the target object based on the transaction trend change characteristics.

[0053] In the embodiments of the present application, the multi-dimensional transaction data at least includes: transaction data of different geographical regions, transaction data of enterprises of different scales, and transaction data of different trading platforms.

[0054] Briefly speaking, the multi-dimensional transaction data in the embodiments of the present application covers transaction data of different geographical regions, transaction data of enterprises of different scales, and transaction data of different trading platforms. The transaction data of different geographical regions can reflect the characteristics of sodium dithionite transactions in each region. For example, in some regions, due to resource distribution or industrial agglomeration, the supply and demand situation and prices will show unique trends. The transaction data of enterprises of different scales shows the different performances of large, medium and small enterprises in terms of transaction quantity, frequency and price based on the differences in their own production, procurement and sales strategies. The transaction data of different trading platforms reflects the differences in trading rules, audience groups, etc. of each platform and the impact on the trading price, activity, etc. of sodium dithionite. These multi-dimensional data comprehensively outline the market transaction situation of sodium dithionite and provide rich information support for subsequent analysis and prediction.

[0055] In the embodiments of the present application, the real-time transaction network model is set with multiple networking modes, and each networking mode is constructed by using the transaction data of the corresponding dimension. Further optionally, the multiple networking modes at least include: geographical region transaction network, enterprise transaction network, platform transaction network, and comprehensive transaction network. Further optionally, the comprehensive transaction network is networked by using the transaction data of at least two dimensions. In the embodiments of the present application, the real-time transaction network model has diversified networking modes, and each mode is built based on the transaction data of the corresponding dimension. These networking modes at least cover the geographical region transaction network, enterprise transaction network, platform transaction network, and comprehensive transaction network.

[0056] Specifically, the geographical region transaction network networks the transaction data of different geographical regions according to their geographical location information. A graph database (such as Neo4j) can be used, regarding each region as a node, and the trade connections between regions as edges, and the weight of the edges can be determined according to indicators such as trade volume and transaction frequency.

[0057] The enterprise transaction network takes enterprises of different scales as nodes and the transaction relationships between enterprises as edges. The attributes of the edges can include information such as transaction amount, transaction time, and transaction frequency. According to the procurement and sales behaviors of enterprises, a directed graph is constructed to reflect the supply and demand relationships between enterprises.

[0058] The platform trading network takes different trading platforms as nodes, and the user flow between platforms, commodity information sharing, etc. are used as the attributes of the edges, reflecting the status and mutual relationships of different platforms in the sodium hydrosulfite trading.

[0059] The comprehensive trading network integrates data from multiple dimensions such as geographical regions, enterprises, and platforms to construct a complex network. For example, by combining geographical region and enterprise data, a trade network among enterprises within the region is formed, and at the same time, considering the role of platforms in it, a multi-dimensional network structure is constructed.

[0060] Exemplarily, the geographical region trading network takes different regions as nodes, and uses trading data of each region, such as trade volume, trading frequency, etc., to construct a network structure reflecting the trade connections between regions, showing the associations and differences in the sodium hydrosulfite trading in different regions. The enterprise trading network takes enterprises of different scales as nodes, and constructs a directed graph based on data such as the trading amount and trading time between enterprises, clearly presenting the supply and demand relationships between enterprises. The platform trading network regards each trading platform as a node, and uses data such as user flow between platforms and commodity information sharing to outline the status and mutual relationships of each platform in the sodium hydrosulfite trading.

[0061] Particularly, the comprehensive trading network constructs a complex structure including the trade network among enterprises within the region by integrating trading data of at least two dimensions, such as combining geographical region and enterprise data, and at the same time considering the role of platforms in it, comprehensively reflecting the complex dynamics of the sodium hydrosulfite market trading, providing strong support for in-depth market analysis.

[0062] For example, the geographical region trading network: It is constructed by taking regions such as East China, South China, and North China in the country as nodes. For instance, the chemical industry in East China is developed and has a large demand for sodium hydrosulfite, forming frequent trade exchanges with North China where resources are abundant. By recording trading data such as the trading volume, transportation routes, and price fluctuations of sodium hydrosulfite between different regions, a geographical region trading network reflecting the supply and demand relationships and logistics directions between regions can be constructed, and it can be intuitively seen that, for example, North China supplies a large amount of sodium hydrosulfite to East China, and the price is affected by transportation costs and local demand.

[0063] The enterprise trading network: Takes enterprises of different scales such as large chemical enterprise A, medium-sized production enterprise B, and small distributor C as nodes. Enterprise A, as the industry leader, purchases raw materials from multiple enterprises for the production of sodium hydrosulfite, and supplies a large amount of goods to enterprises B and C. Enterprise C then sells the products to some small end-users. By constructing a network based on data such as purchase orders and sales contracts between enterprises, the core position of enterprise A in the industrial chain and the upstream and downstream supply relationships between enterprises can be presented.

[0064] Platform trading network: Nodes include chemical product trading platform D, comprehensive e-commerce platform E, etc. Platform D mainly focuses on the trading of chemical raw materials, with a large number of professional chemical enterprises settled in. Platform E targets a wider range of commodity fields, but also has some transactions of sodium hydrosulfite. By analyzing the traffic data, trading activity, user reviews, etc. of each platform, a platform trading network can be constructed, and it can be found that Platform D is more professional and concentrated in the trading of sodium hydrosulfite, while the transactions on Platform E are more dispersed, mainly small orders.

[0065] Comprehensive trading network: Constructed by integrating data from multiple dimensions such as geographical regions, enterprises, and platforms. For example, integrating the data of enterprise A in the East China region, enterprise B in the South China region with that of Platform D and Platform E. It can be found that enterprise A has trading contacts with enterprise B in the South China region through Platform D, and also has a small number of transactions with some small enterprises on Platform E, presenting a complex network structure where enterprises in different regions conduct transactions through different platforms, comprehensively showing the trading relationships and dynamics of sodium hydrosulfite in different dimensions.

[0066] In the technical solution of this application, when obtaining the trading data of sodium hydrosulfite, the conventional practice of relying only on a single type of data is abandoned. Data on the supply and demand, prices, etc. of sodium hydrosulfite are collected from different geographical regions, such as the concentrated coastal chemical industry areas in the east and the inland resource production areas; the procurement, sales, and inventory data of large, medium, and small enterprises are integrated; at the same time, the trading information of various online and offline trading platforms is included. Combining the strong demand data in a certain coastal area, the expanded production supply data of large production enterprises, and the price fluctuation data of popular online platforms can present the market situation in all aspects and provide sufficient basis for accurate analysis.

[0067] When the target object raises a query demand, detailed information such as its geographical location, user type (such as manufacturer, distributor, retailer), size and nature of the organization, and available trading platforms is collected. For example, if the target object is a small food processing enterprise in the west, through these attribute information, its position in the regional supply and demand, enterprise transaction level and applicable platform can be accurately located, so as to customize its exclusive price trend analysis to meet its unique needs. Construct a geographical regional transaction network, with each region as a node, set edges according to the intensity of trade exchanges, and intuitively show the flow of supply and demand between regions. The enterprise transaction network uses enterprises as nodes and sets edges according to the upstream and downstream relationships of the supply chain to highlight the commercial connection between enterprises. The platform transaction network uses platforms as nodes and sets edges according to the flow and information interaction between platforms to reveal the influence of the platform. The comprehensive transaction network integrates multiple dimensions, such as combining the data of enterprises in the central and western regions and online and offline platforms to build a complex network and fully display the market structure. The model continuously connects to real-time monitoring data streams, such as obtaining the latest transaction prices, enterprise inventory changes, and platform transaction order volume in each region at regular intervals. When there is an unexpected market situation, such as transportation obstruction in a certain area, the model can quickly capture data changes, update price forecasts, and provide a basis for market participants to adjust production, procurement, and sales strategies in a timely manner. Use the change trend prediction model to conduct in-depth analysis of the real-time transaction network model. For example, through analysis, it was found that when enterprises in a specific region trade on a specific trading platform, price fluctuations are related to platform promotion activities and regional seasonal demand. This potential relationship is difficult to discover with traditional methods, but it can provide key clues for market forecasting. For production companies, the model can provide decision-making suggestions such as expanding or reducing production capacity and adjusting procurement channels based on the supply and demand in their region, the dynamics of cooperative enterprises, and the competitive situation of applicable platforms; for traders, it can provide strategies for optimizing procurement and sales paths based on price differences on different platforms and changes in regional supply and demand, helping various market players stand out in the competition.

[0068] In the embodiment of the present application, the attribute information includes at least one of the following: the geographical location of the target object, the user type of the target object, the organization to which the target object belongs, and the trading platform available to the target object.

[0069] In the embodiments of the present application, the attribute information includes multiple key dimensions, which are used to accurately depict the characteristics of the target object in the sodium dithionite trading market. The geographical location of the target object is crucial. There are differences in supply and demand conditions, transportation costs, industrial distributions, etc. in different regions, which will greatly affect the trading price and market dynamics of sodium dithionite. For example, in regions where production enterprises are concentrated, the price may show a specific trend due to competition and supply conditions. The user type to which the target object belongs cannot be ignored either. Whether it is a production enterprise, a trader, or an end consumer, their demands and trading models are different. For example, production enterprises focus on the stable supply and price of raw materials, while traders value price fluctuations and profit margins. The institution to which the target object belongs. Different institutions have different influences in the market, resource integration capabilities, and procurement and sales strategies, and there will also be differences in the trading behaviors and key concerns of sodium dithionite. In addition, for the trading platforms available to the target object, the trading rules, user groups, trading activity levels, and price formation mechanisms of different platforms are different, which determines the channels and price ranges for the target object to obtain products, and thus affects its trading decisions and market participation methods. These attribute information combined can provide key basis for subsequent accurate analysis and prediction.

[0070] As an optional embodiment, multi-dimensional trading data of sodium dithionite is obtained in 101. The multi-dimensional trading data at least includes: trading data in different geographical regions, trading data of enterprises of different scales, and trading data of different trading platforms. The following is a specific introduction to different types of data in the multi-dimensional trading data of sodium dithionite:

[0071] Production location data: For sodium dithionite, in chemical industry concentrated regions such as Shandong and Jiangsu, the production volume and inventory data of production enterprises are crucial. For example, the monthly production volume data of enterprises in a large chemical industrial park in Shandong, and information such as the shipment volume of these enterprises to surrounding regions can reflect the supply capacity of the production location.

[0072] Consumption location data: In developed manufacturing regions such as Guangdong and Zhejiang, which are the main consumption locations of sodium dithionite, data such as the procurement volume, consumption volume, and inventory levels of local enterprises can reflect the demand situation in the consumption location. For example, the quarterly change in the procurement volume of sodium dithionite by a printing and dyeing enterprise cluster in Guangdong.

[0073] Import and export data: The import and export data of sodium dithionite in coastal port cities such as Shanghai and Tianjin, including the source countries of imports, destinations of exports, import and export volumes, and prices, can reflect the impact of the international market on the domestic market and the status of the domestic market in the international arena.

[0074] Large enterprise data: Large chemical production enterprises such as Jineng Technology, their production plans, sales channels, and long-term contract data with upstream and downstream enterprises have a significant impact on the overall supply and demand balance in the market. For example, the annual supply contract volume between Jineng Technology and large paper-making enterprises.

[0075] Data of medium-sized enterprises: Medium-sized enterprises play a connecting role in the market. Their expansion plans, the impact of technological transformation on production capacity, and transaction data in regional markets, such as changes in market share of a medium-sized enterprise in North China, are also important analysis dimensions.

[0076] Small business data: Although small businesses are small in size, they are numerous, and their flexible market strategies, sensitivity to price fluctuations, and other data cannot be ignored. For example, the purchase frequency and single purchase volume changes of small printing and dyeing enterprises when raw material prices rise.

[0077] Offline traditional trading platform data: data on the on-site trading volume, trading price, information on both parties of sodium dithionite on offline platforms such as chemical product trading markets and industry exhibitions can reflect the activity and trading patterns of traditional offline transactions. For example, the transaction price range of sodium dithionite in different time periods in a chemical trading market.

[0078] Online e-commerce platform data: data such as product listing prices, transaction volumes, page views, and user inquiry records on online platforms such as Mobai and Gaide Chemical Network can reflect online transaction trends and market popularity. For example, the transaction volume growth of sodium dithionite on the Mobai platform during the promotion period.

[0079] Enterprise-owned platform data: Transaction data on the transaction platforms built by some large enterprises, including direct transaction records between enterprises and customers, customized service order data, etc., can reflect the enterprise's own sales strategy and customer relationship management.

[0080] As an optional embodiment, the establishment of a real-time transaction network model of sodium dithionite based on the multi-dimensional transaction data in 102 can be implemented as follows:

[0081] Separating the multi-dimensional transaction data according to different dimensions to obtain transaction data under different dimensions; wherein the different dimensions at least include: geographical area, enterprise size, enterprise type, transaction platform, and transaction time;

[0082] Based on the preset networking conditions of different dimensions, establish the network model space corresponding to different dimensions;

[0083] Projecting the transaction data in different dimensions into the corresponding network model spaces as corresponding nodes to obtain first real-time transaction network models corresponding to different dimensions;

[0084] Adaptively learn the transaction data under different dimensions to obtain the connection relationship between the nodes in the first real-time transaction network model; the meaning represented by the connection relationship between the nodes in the first real-time transaction network model is associated with the data type of the corresponding dimension.

[0085] Exemplarily, assume there is a large amount of sodium dithionite transaction data. First, separate these data according to different dimensions. For example, from the geographical region dimension, the data are divided into transaction data of regions such as East China, South China, and North China; from the enterprise scale dimension, they are divided into data of large, medium, and small enterprises; from the enterprise type dimension, they are divided into data of production enterprises, trading enterprises, etc.; from the transaction platform dimension, they are divided into data of online e-commerce platforms and offline traditional markets; from the transaction time dimension, the data are divided by quarter and month. Then, establish a network model space according to the preset networking conditions of different dimensions. For the geographical region dimension, taking regions as nodes, if there is frequent sodium dithionite trade between two regions, a connection is established. Project the transaction data of the East China region into the corresponding geographical region network model space as a node. In this way, the East China region becomes a node in the geographical region network. Similarly, other regions are also like this, forming the first real-time transaction network model corresponding to the geographical region. Conduct adaptive learning on the data of different dimensions to determine the node connection relationship. In the enterprise scale dimension, between large enterprises and medium-sized enterprises that supply them with raw materials and small enterprises that purchase their products, by analyzing the transaction data, it is determined that there is a business connection relationship. This connection relationship may represent the logistics transportation relationship between regions in the geographical region network; in the enterprise scale network, it represents the upstream and downstream relationship of the supply chain, clearly presenting the internal connections of sodium dithionite transactions under each dimension and helping to comprehensively understand the market transaction dynamics.

[0086] Further optionally, in the above steps, conducting adaptive learning on the transaction data under different dimensions to obtain the connection relationship between each node in the first real-time transaction network model can be implemented as:

[0087] Set each geographical region in the geographical region transaction network as a node, and set the trade relationship between different geographical regions as an edge; set the weight of each edge according to the trade volume and transaction frequency of the geographical region where it is located; where the weight w of the edge between the i-th node and the j-th node ij is expressed as: α is an adjustable parameter, α is not less than 0 and not greater than 1, V ij is the trade volume between the region i corresponding to the i-th node and the region j corresponding to the j-th node, F ij is the transaction frequency between the region i corresponding to the i-th node and the region j corresponding to the j-th node, represents the sum of the trade volume V ik between region i and other regions k, and F is the transaction frequency between region i and other regions k ikThe sum, where k is the summation index, N is the total number of geographical regions, and the value range of k is from 1 to N. When it is 1, the edge weight is mainly determined by the trade volume; when it is 0, the edge weight is mainly determined by the trading frequency.

[0088] The function of the above formula is to quantify the weights of the edges between nodes in the geographical region trading network, so as to more accurately reflect the tightness of the trade relationship between different geographical regions. Through this formula, two key factors, namely the trade volume and trading frequency between regions, are comprehensively considered. The trade volume reflects the scale of the sodium dithionite trade between regions, and the trading frequency reflects the activity of the transactions. Incorporating these two factors into the weight calculation can more comprehensively depict the trade connections between regions. For example, for the situation where the trade volume between two geographical regions is large but the trading frequency is low, compared with the situation where the trade volume is small but the trading frequency is high, different weights will be obtained through formula calculation, thus reflecting different connection strengths in the network model.

[0089] At the same time, in the formula, by summing up the trade volume and trading frequency between region i and all other regions and introducing adjustable parameters, the calculation of the weights can not only highlight the relationships between specific regions but also be measured in the context of the overall network. This enables the model to flexibly adjust the focus of the weights according to the actual situation, making the connection relationships between nodes in the geographical region trading network more in line with the actual trade situation, and providing a more accurate and effective basis for analyzing the trading patterns and trends of sodium dithionite between different geographical regions.

[0090] Further optionally, in the above steps, adaptive learning of the trading data in different dimensions to obtain the connection relationships between each node in the first real-time trading network model can be implemented as follows:

[0091] Set each enterprise in the enterprise trading network as a node, and the trading relationships between different enterprises as edges; based on the trading amount, trading time, and trading frequency between each enterprise and other enterprises, set the weights of the edges corresponding to each enterprise; where the weight w of the edge between the m-th node and the n-th node mn is expressed as: where β1, β2, and β3 are adjustable parameters, and the sum of β1, β2, and β3 is 1, A mn is the trading amount between the enterprise m corresponding to the m-th node and the enterprise n corresponding to the n-th node, T mn is the reciprocal of the time from the most recent trading time to the current time, F mn is the trading frequency between the enterprise m corresponding to the m-th node and the enterprise n corresponding to the n-th node, represents the sum of the trading amounts A between enterprise m and other enterprises k mk and represents the reciprocal of the trading time T between enterprise m and other enterprises kmk The sum of is the transaction frequency F between enterprise m and other enterprise k mk The sum, where k is the summation index, M is the total number of enterprises, and the value range of k is from 1 to M.

[0092] The function of the above formula is to calculate the weight of the edge between nodes in the enterprise transaction network to reflect the trading tightness and importance between enterprises. It comprehensively considers three important factors: transaction amount, transaction time, and transaction frequency, making the weight of the edge more representative.

[0093] The transaction amount reflects the scale of business transactions between enterprises and has a direct impact on the weight. The reciprocal of the transaction time converts it into the sensitivity to time, and the weight of transactions closer to the current time is higher, reflecting the timeliness of transactions. The transaction frequency reflects the activity of transactions between enterprises. By combining these three factors through adjustable parameters and adjusting their proportions in the weight according to different needs, the weight of the edge can be flexibly adjusted to more accurately describe the trade relationship between enterprises. This calculation method helps to comprehensively and meticulously display the mutual connections between enterprises in the sodium dithionite transaction, providing a more realistic data basis for subsequent market analysis and prediction, and enabling the network model to better present the actual situation and importance of transactions between enterprises.

[0094] Further optionally, in the above steps, adaptive learning of transaction data in different dimensions to obtain the connection relationship between each node in the first real-time transaction network model can be implemented as:

[0095] Set each trading entity in different trading platforms in the platform trading network as a node, and the trading relationship between each trading entity in different trading platforms as an edge; set the weight of each edge according to the user flow and commodity information sharing between each trading entity; where the weight w of the edge between the p-th node and the q-th node in the current trading platform pq is expressed as: where γ is an adjustable parameter, γ is not less than 0 and not greater than 1, U pq is the user flow between the trading entity p corresponding to the p-th node and the trading entity q corresponding to the q-th node, S pq is the degree of commodity information sharing between the trading entity p corresponding to the p-th node and the trading entity q corresponding to the q-th node, represents the sum of the user flow U between trading entity p and other entity k pk The sum of is the degree of commodity information sharing S between trading entity p and other entity k pk The sum, where k is the summation index, L is the total number of entities in the current trading platform, the value range of k is from 1 to L, λt is the time decay factor at the current moment t, R p and R q are the platform credit ratings corresponding to the trading entities p and q respectively.

[0096] The function of the above formula is to accurately calculate the weight of the edge between nodes in the platform trading network, comprehensively reflecting the closeness of the relationship between different trading entities. It comprehensively considers the user flow volume, the degree of commodity information sharing, the platform credit rating, and the time decay factor. The user flow volume and the degree of commodity information sharing reflect the interaction activity and information transmission intensity between trading entities. The time decay factor makes recent information more important, which conforms to the dynamic characteristics of the market. The introduction of the platform credit rating indicates that the platform's reputation will affect the importance of the relationship between trading entities. The adjustable parameter can adjust the weight ratio of the user flow volume and the degree of commodity information sharing according to the actual situation, enabling the weight of the edge to vary flexibly according to different scenarios, so as to more accurately depict the relationship between various trading entities in the trading platform, providing more reasonable data support for accurately analyzing the trading situation of sodium dithionite in the platform and subsequent market forecasts.

[0097] As an optional embodiment, in 103, a query instruction of a target object is received to obtain the attribute information of the target object. The attribute information includes at least one of the following: the geographical location where the target object is located, the user type to which the target object belongs, the institution to which the target object belongs, and the trading platform available to the target object.

[0098] In step 103, receiving the query instruction of the target object is to accurately obtain its attribute information. The geographical location where the target object is located is crucial. For example, whether it is located in the coastal chemical industry belt or the inland area will affect its supply and demand of sodium dithionite and price sensitivity. The user type to which the target object belongs is also important. If it is a production enterprise, its focus may be on raw material procurement costs and supply stability. If it is a trader, it is more concerned about price fluctuations and market price differences. If it is an end user, it pays more attention to product quality and usage effects. The information of the institution to which the target object belongs can reflect the status and influence of the object in the market, and different institutions may have different procurement and sales strategies. And the available trading platform will determine its channel for obtaining sodium dithionite. For example, using an online platform may have richer information but greater competition, while using an offline platform may focus more on long-term stable cooperation relationships. These attribute information are very important for subsequent personalized services and market analysis, and can more targeted provide services such as price trend prediction of sodium dithionite for the target object.

[0099] Furthermore, when the attribute information of the target object changes, a real-time monitoring mechanism can be established. For example, data interfaces, regular web crawlers, etc. can be used to regularly obtain the latest attribute information of the target object. An information push mechanism can also be set up to enable the target object to actively inform when its attribute information changes. An early warning system can be established to send timely reminders when key attribute information changes are detected, so that relevant personnel can quickly respond. If the target object relocates to a new area, it is necessary to re-evaluate the local market environment, logistics distribution, etc., and adjust the supply channels or recommend new partners. When a manufacturing enterprise transforms into a trading company, the service content should change from providing production technical support to market trend analysis, trade channel expansion, etc. If the institution to which the target object belongs undergoes mergers, splits, etc., the service plan should be adjusted according to the characteristics and needs of the new institution, such as providing services suitable for the scale and business scope of the new institution. If the target object starts using a new trading platform, it is necessary to provide guidance on using the new platform, interpretation of trading rules, and exclusive preferential information for the new platform. Further optionally, communicate with the target object in a timely manner to understand its needs and expectations for the new service, and further optimize the service content and methods based on the feedback. For changes in attribute information, provide personalized information push to the target object, such as sending introductions and usage instructions of the new service content, to ensure that it can quickly understand and adapt to the new service.

[0100] As an alternative embodiment, assume that the change trend prediction model at least includes the following structures: an input layer, a position encoding layer, an adaptive prediction layer, and an output layer. Based on this, in 104, by using the change trend prediction model to analyze the real-time monitoring data stream corresponding to the attribute information in the real-time trading network model, the change trend characteristics of sodium dithionite trading matching the target object can be obtained, which can be implemented as:

[0101] Through the input layer, extract the real-time monitoring data stream corresponding to the attribute information from the real-time trading network model;

[0102] Through the position encoding layer, perform position encoding on the positions of each node in the real-time monitoring data stream to obtain the position encoding information corresponding to each node;

[0103] Through the adaptive prediction layer, based on the position encoding information, capture the parallel dependency relationships of the first attribute feature information corresponding to each node and the second attribute feature information between each node, and adaptively predict the trading trend characteristics corresponding to each node based on the captured dependency relationships;

[0104] Through the output layer, fuse the trading trend characteristics corresponding to each node into the trading trend change characteristics.

[0105] In this alternative embodiment, the input layer is first used to extract the real-time monitoring data stream corresponding to the target object attribute information from the real-time trading network model. Then, the position encoding layer encodes the positions of the respective nodes in these data streams to obtain position encoding information, providing a basis for position information for subsequent processing. Next, the adaptive prediction layer, based on the position encoding information, simultaneously captures the dependency relationships of the first attribute feature information of each node and the second attribute feature information between the nodes in parallel. On this basis, it adaptively predicts the trading trend features of each node using the captured dependency relationships. Finally, the output layer fuses the trading trend features of these nodes to ultimately obtain the sodium dithionite trading trend change features matching the target object, thereby connecting the entire process from data extraction to feature processing and prediction, realizing the prediction of the sodium dithionite trading trend, and providing key information for subsequent decision-making and analysis.

[0106] This process can, based on the attribute information of the target object, combined with the real-time monitoring data stream, make full use of the structure of the network model, give play to the functions of each layer, effectively mine the potential information and relationships in the data, and predict the trading trend change features that conform to the actual situation.

[0107] For example, combined with novel neural network algorithms, such as algorithms based on the Transformer architecture, the change trend prediction model can be implemented as follows. The input layer is responsible for receiving the real-time monitoring data stream corresponding to the target object attribute information from the real-time trading network model. These data may include various data related to sodium dithionite trading under different dimensions such as different geographical regions, enterprise scales, trading platforms, etc., such as trading volume, price, trading frequency, etc.

[0108] The position encoding layer draws on the idea of the Transformer architecture. Since it is difficult for the neural network itself to capture the position information of the data, this layer assigns position encoding information to the respective nodes in the real-time monitoring data stream through a specific algorithm. For example, by using a combination of sine and cosine functions, the position information is encoded into a vector with the same dimension as the data features, enabling the model to perceive the sequential relationship of the data, which is crucial for analyzing the change trend of trading data over time or other sequential dimensions.

[0109] The adaptive prediction layer is the core part of the model. Also referring to the Transformer architecture, it uses the multi-head self-attention mechanism to capture the dependency relationships between the first attribute feature information corresponding to each node (such as features like the trading price and trading volume of the node itself) and the second attribute feature information between each node (such as features like the trading relationship and information sharing degree between different nodes) in parallel. Through the parallel computing of multiple heads, the multi-head self-attention mechanism can capture complex patterns in the data from different perspectives simultaneously. Each head assigns different attention weights to the data at different positions by calculating the relationships between the query, key, and value matrices, thus highlighting the information important for predicting changes in the trading trend. After capturing the dependency relationships, these information are further processed and integrated through a feed-forward neural network to adaptively predict the trading trend features corresponding to each node. The feed-forward neural network consists of multiple fully connected layers with activation functions interspersed, which can perform non-linear transformations on the data and enhance the expressive power of the model.

[0110] The output layer fuses the trading trend features corresponding to each node. A simple weighted summation method can be used, or a more complex neural network layer (such as a multi-layer perceptron) can be used to learn how to effectively fuse these features into the final trading trend change features. The finally obtained trading trend change features will reflect the trading trend of sodium dithionite in the context related to the attributes of the target object, providing valuable decision-making basis for market participants. During the training process of the entire model, the model parameters are continuously adjusted through the backpropagation algorithm to minimize the error between the prediction result and the actual trading trend, thereby improving the prediction accuracy of the model.

[0111] As an optional embodiment, in 105, the corresponding sodium dithionite price change trend information is generated based on the trading trend change features and presented to the target object.

[0112] For example, after obtaining the trading trend change features through the previous steps, it enters the key link of generating and presenting the sodium dithionite price change trend information. First, the system deeply analyzes and transforms according to the obtained trading trend change features by using specific algorithms and models. This may involve complex calculations of the correlations between multiple factors such as trading volume, trading frequency, and market supply and demand relationships with the price, converting the abstract trading trend features into intuitive price change trend data. Subsequently, the system will present these price change trend data in an easy-to-understand visual form, such as line charts, bar charts, or dynamic charts, so that the target object can clearly see the rising, falling, or fluctuating trends of the sodium dithionite price in different time dimensions.

[0113] In practical applications, further optionally, the display method may further include detailed written explanations elaborating on the reasons for price changes, influencing factors, and possible future trends. During the display process, the system will consider the usage habits and preferences of the target object and perform precise push through the specified platforms or channels. Whether it is professional analysis software on the computer side or a convenient application on the mobile side, it ensures that the target object can receive the price change trend information of sodium dithionite in a timely and accurate manner, thereby providing strong support for its subsequent production, procurement, sales, and other decisions.

[0114] Further optionally, in 105, after generating and displaying the corresponding price change trend information of sodium dithionite to the target object based on the transaction trend change characteristics, it is also possible to identify the abnormal feature points in the transaction trend change characteristics; based on the abnormal feature points, retrieve the corresponding historical transaction data from the real-time transaction network model; determine the candidate abnormal types corresponding to the abnormal feature points based on the historical transaction data, and generate corresponding price warning information, so that the target object can learn about the corresponding potential transaction risks based on the price warning information.

[0115] For example, assume that the transaction trend change characteristics of sodium dithionite have been obtained through a change trend prediction model and the price change trend information has been displayed to the target object. After that, the system starts to automatically identify abnormal feature points. For example, in the trend chart of price fluctuations, a point where the price suddenly rises significantly and the trading volume surges may be identified as an abnormal feature point. Then, the system quickly retrieves the corresponding historical transaction data from the real-time transaction network model. Through data tracing, it is found that when this abnormal point occurred, there was a large-scale centralized procurement behavior among several large enterprises in a specific region, and the inventory of suppliers in this region decreased sharply in the short term. Based on these historical transaction data, the system determines the candidate abnormal types corresponding to the abnormal feature points. After analysis, it may be due to problems with the supply capacity of suppliers, resulting in a short-term supply shortage and thus triggering abnormal price fluctuations. Therefore, the abnormal type is determined to be "supply shortage risk". Finally, the system generates corresponding price warning information, such as "Recently, due to supply shortages in a certain region in the sodium dithionite market, prices may continue to rise, posing a risk of a significant increase in procurement costs", and pushes it to the target object. In this way, the target object can, based on this price warning information, be aware of potential transaction risks in advance, thereby adjusting the procurement plan, seeking alternative suppliers, or taking other countermeasures to reduce the possible losses.

[0116] In the embodiments of the present application, by integrating multi-dimensional data of different geographical regions, enterprises of different scales, and trading platforms, the overall market picture is presented from multiple perspectives, price influencing factors are comprehensively captured, and the prediction accuracy is significantly improved. Moreover, based on the attribute information such as the geographical location and user type of the target object, personalized price trend information is provided to meet the specific needs of each entity and assist them in making favorable decisions. Furthermore, by establishing a real-time trading network model and analyzing the real-time monitored data stream, the market dynamics are tracked in a timely manner, enabling price prediction to reflect market changes in real time and facilitating participants to adjust their strategies in a timely manner. Finally, the change trend prediction model can also be used to analyze the real-time trading network model to explore the potential relationships between data in different dimensions and discover price influencing factors and patterns that are difficult to detect by traditional methods, such as specific price change rules between different elements in the comprehensive trading network. The embodiments of the present application can achieve the price trend prediction of sodium dithionite in the case of data shortage and poor accuracy, and improve the reliability and timeliness of the sodium dithionite price prediction.

[0117] In another embodiment of the present application, a sodium dithionite price trend prediction system is further provided. Refer to Figure 2 As described, the sodium dithionite price trend prediction system includes the following units:

[0118] The acquisition unit is configured to obtain multi-dimensional trading data of sodium dithionite; the multi-dimensional trading data at least includes: trading data of different geographical regions, trading data of enterprises of different scales, and trading data of different trading platforms;

[0119] The networking unit is configured to establish a real-time trading network model of sodium dithionite based on the multi-dimensional trading data; the real-time trading network model is provided with multiple networking modes, and each networking mode is constructed by using the trading data of the corresponding dimension; the multiple networking modes at least include: geographical region trading network, enterprise trading network, platform trading network, and comprehensive trading network; the comprehensive trading network is formed by networking the trading data of at least two dimensions;

[0120] The interaction unit is configured to receive a query instruction of the target object to obtain the attribute information of the target object; the attribute information at least includes one of the following: the geographical location where the target object is located, the user type to which the target object belongs, the institution to which the target object belongs, and the trading platform available to the target object;

[0121] The prediction unit is configured to analyze the real-time monitored data stream corresponding to the attribute information in the real-time trading network model by using a change trend prediction model to obtain the change characteristics of the sodium dithionite trading trend matched by the target object;

[0122] The interaction unit is further configured to generate and display the corresponding sodium dithionite price change trend information to the target object based on the trading trend change characteristics.

[0123] Further optionally, the networking unit establishes a real-time trading network model of sodium dithionite based on the multi-dimensional trading data, and is configured to:

[0124] Separate the multi-dimensional trading data according to different dimensions to obtain trading data under different dimensions; wherein, the different dimensions at least include: geographical region, enterprise scale, enterprise type, trading platform, trading time;

[0125] Based on the preset networking conditions of different dimensions, establish network model spaces corresponding to different dimensions;

[0126] Project the trading data under different dimensions into their respective corresponding network model spaces as corresponding nodes to obtain a first real-time trading network model corresponding to different dimensions;

[0127] Perform adaptive learning on the trading data under different dimensions to obtain the connection relationships between the nodes in the first real-time trading network model; the meaning represented by the connection relationships between the nodes in the first real-time trading network model is associated with the data types of the corresponding dimensions.

[0128] Further optionally, the networking unit performs adaptive learning on the trading data under different dimensions to obtain the connection relationships between the nodes in the first real-time trading network model, and is configured to:

[0129] Set each geographical region in the geographical region trading network as a node, and set the trade relationships between different geographical regions as edges;

[0130] Set the weight of each edge according to the trade volume and trading frequency of the geographical region where it is located; wherein, the weight w of the edge between the i-th node and the j-th node ij is expressed as: α is an adjustable parameter, α is not less than 0 and not greater than 1, V ij is the trade volume between the region i corresponding to the i-th node and the region j corresponding to the j-th node, F ij is the trading frequency between the region i corresponding to the i-th node and the region j corresponding to the j-th node, represents the sum of the trade volumes V ik between the region i and other regions k, is the sum of the trading frequencies F ik between the region i and other regions k, k is the summation index, N is the total number of geographical regions, and the value range of k is from 1 to N.

[0131] Further optionally, the networking unit performs adaptive learning on the trading data under different dimensions to obtain the connection relationships between the nodes in the first real-time trading network model, and is configured to:

[0132] Each enterprise in the enterprise transaction network is set as a node, and the transaction relationships between different enterprises are used as edges;

[0133] Based on the transaction amount, transaction time, and transaction frequency between each enterprise and other enterprises, the weight of the edge corresponding to each enterprise is set;

[0134] Among them, the weight w of the edge between the m-th node and the n-th node mn is expressed as:

[0135] Among them, β1, β2, and β3 are adjustable parameters, and the sum of β1, β2, and β3 is 1. A mn is the transaction amount between the enterprise m corresponding to the m-th node and the enterprise n corresponding to the n-th node, T mn is the reciprocal of the time from the most recent transaction time to the current time, F mn is the transaction frequency between the enterprise m corresponding to the m-th node and the enterprise n corresponding to the n-th node, represents the sum of the transaction amounts A between the enterprise m and other enterprises k mk and, represents the sum of the reciprocals of the transaction times T between the enterprise m and other enterprises k mk and, is the sum of the transaction frequencies F between the enterprise m and other enterprises k mk and, k is the summation index, M is the total number of enterprises, and the value range of k is from 1 to M.

[0136] Further optionally, the networking unit adaptively learns the transaction data in different dimensions to obtain the connection relationships between the nodes in the first real-time transaction network model, and is configured as:

[0137] Each transaction subject in different trading platforms in the platform transaction network is set as a node, and the transaction relationships between the transaction subjects in different trading platforms are used as edges;

[0138] The weight of each edge is set according to the user flow and commodity information sharing between the transaction subjects;

[0139] Among them, the weight w of the edge between the p-th node and the q-th node in the current trading platform pq is expressed as:

[0140] Among them, γ is an adjustable parameter, γ is not less than 0 and not greater than 1, U pq is the user flow volume between the transaction subject p corresponding to the p-th node and the transaction subject q corresponding to the q-th node, S pqis the degree of commodity information sharing between the trading entity p corresponding to the p-th node and the trading entity q corresponding to the q-th node. represents the user flow volume U between the trading entity p and other entity k pk sum is the degree of commodity information sharing S between the trading entity p and other entity k pk sum, k is the summation index, L is the total number of entities in the current trading platform, the value range of k is from 1 to L, λ t is the time decay factor at the current moment t, R p and R q are respectively the platform reputation ratings corresponding to the trading entity p and the trading entity q.

[0141] Further optionally, the change trend prediction model at least includes the following structures: an input layer, a position encoding layer, an adaptive prediction layer, and an output layer;

[0142] The prediction unit, using the change trend prediction model, analyzes the real-time monitoring data stream corresponding to the attribute information in the real-time trading network model to obtain the change trend characteristics of the sodium dithionite transaction matching the target object, and is configured to:

[0143] Extract the real-time monitoring data stream corresponding to the attribute information from the real-time trading network model through the input layer;

[0144] Through the position encoding layer, perform position encoding on the positions of each node in the real-time monitoring data stream to obtain the position encoding information corresponding to each node;

[0145] Through the adaptive prediction layer, based on the position encoding information, capture the parallel dependency relationships of the first attribute feature information corresponding to each node and the second attribute feature information between each node, and adaptively predict the transaction trend characteristics corresponding to each node based on the captured dependencies;

[0146] Through the output layer, fuse the transaction trend characteristics corresponding to each node into the transaction trend change characteristics.

[0147] Further optionally, after the interaction unit generates and displays the corresponding sodium dithionite price change trend information to the target object based on the transaction trend change characteristics, it is further configured to:

[0148] Identify the abnormal feature points in the transaction trend change characteristics;

[0149] Based on the abnormal feature points, retrieve the corresponding historical transaction data from the real-time trading network model;

[0150] Determine a candidate abnormal type corresponding to the abnormal feature point based on the historical transaction data, and generate a corresponding price warning message, so that the target object can learn the corresponding potential transaction risk based on the price warning message.

[0151] The system can implement various steps in the above method embodiments, which will not be elaborated here for the time being.

[0152] In the embodiments of the present application, the price trend prediction of sodium hydrosulfite is realized under the condition of data loss and poor accuracy, improving the reliability and timeliness of the sodium hydrosulfite price prediction.

[0153] Please refer to Figure 3 , Figure 3 which is a schematic diagram of an embodiment of an electronic device provided by an embodiment of the present invention. As Figure 3 shown, an embodiment of the present invention provides an electronic device 500, including a memory 510, a processor 5102, and a computer program 511 stored on the memory 510 and executable on the processor 5102. When the processor 5102 executes the computer program 511, the following steps are implemented: obtaining multi-dimensional transaction data of sodium hydrosulfite; the multi-dimensional transaction data at least includes: transaction data of different geographical regions, transaction data of different scale enterprises, and transaction data of different transaction platforms; establishing a real-time transaction network model of sodium hydrosulfite based on the multi-dimensional transaction data; the real-time transaction network model is provided with a variety of networking modes, and each networking mode is constructed by using transaction data of the corresponding dimension; the variety of networking modes at least includes: geographical region transaction network, enterprise transaction network, platform transaction network, and comprehensive transaction network; the comprehensive transaction network is formed by networking transaction data of at least two dimensions; receiving a query instruction of a target object to obtain attribute information of the target object; the attribute information at least includes one of the following: the geographical location where the target object is located, the user type to which the target object belongs, the institution to which the target object belongs, and the transaction platform available to the target object; using a change trend prediction model to analyze the real-time monitoring data stream corresponding to the attribute information in the real-time transaction network model to obtain the change characteristics of the sodium hydrosulfite transaction trend matched by the target object; generating and displaying corresponding sodium hydrosulfite price change trend information to the target object based on the transaction trend change characteristics.

[0154] Please refer to Figure 4 , Figure 4 which is a schematic diagram of an embodiment of a computer-readable storage medium provided by an embodiment of the present invention. As Figure 4As shown in the figure, this embodiment provides a computer-readable storage medium 600, on which a computer program 611 is stored. When the computer program 611 is executed by a processor, the following steps are implemented: obtaining multi-dimensional transaction data of sodium dithionite; the multi-dimensional transaction data at least includes: transaction data in different geographical regions, transaction data of enterprises of different scales, and transaction data of different trading platforms; establishing a real-time transaction network model of sodium dithionite based on the multi-dimensional transaction data; the real-time transaction network model is provided with multiple networking modes, and each networking mode is constructed by using transaction data of the corresponding dimension; the multiple networking modes at least include: geographical region transaction network, enterprise transaction network, platform transaction network, and comprehensive transaction network; the comprehensive transaction network is networked by using transaction data of at least two dimensions; receiving a query instruction of a target object to obtain attribute information of the target object; the attribute information at least includes one of the following: the geographical location where the target object is located, the user type to which the target object belongs, the institution to which the target object belongs, and the trading platform available to the target object; using a change trend prediction model to analyze the real-time monitoring data stream corresponding to the attribute information in the real-time transaction network model to obtain the change trend characteristics of the sodium dithionite transaction matching the target object; generating and displaying corresponding sodium dithionite price change trend information to the target object based on the transaction trend change characteristics.

[0155] It should be noted that in the above embodiments, the descriptions of each embodiment have their own emphases. For the parts not described in detail in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0156] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0157] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate for implementing in the process Figure 1 one process or multiple processes and / or blocks Figure 1means for the functions specified in one or more boxes.

[0158] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufacture including an instruction means that implements the functions specified in one Figure 1 one process or more processes and / or boxes Figure 1 means for the functions specified in one box or more boxes.

[0159] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they learn the basic creative concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications that fall within the scope of the present invention.

[0160] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. A method for predicting the price trend of sodium dithionite, characterized in that: The method comprises: Acquire multi-dimensional transaction data of sodium dithionite; the multi-dimensional transaction data at least includes: transaction data of different geographical regions, transaction data of enterprises of different sizes, and transaction data of different transaction platforms; A real-time transaction network model of sodium dithionite is established based on the multi-dimensional transaction data; the real-time transaction network model is provided with a plurality of networking modes, each of which is constructed by using transaction data of a corresponding dimension; the plurality of networking modes at least include: a geographical area transaction network, an enterprise transaction network, a platform transaction network, and a comprehensive transaction network; the comprehensive transaction network is obtained by using transaction data of at least two dimensions to form a network; Receive a query instruction of a target object to obtain attribute information of the target object; the attribute information includes at least one of the following: the geographical location of the target object, the user type of the target object, the organization to which the target object belongs, and the trading platform available to the target object; Using a change trend prediction model, analyzing the real-time monitoring data stream corresponding to the attribute information in the real-time transaction network model, so as to obtain the change trend characteristics of sodium dithionite transaction matching the target object; Based on the transaction trend change characteristics, corresponding sodium dithionite price change trend information is generated and displayed to the target object.

2. The method for predicting sodium dithionite price trend according to claim 1, characterized in that: The method of establishing a real-time transaction network model of sodium dithionite based on the multi-dimensional transaction data includes: Separating the multi-dimensional transaction data according to different dimensions to obtain transaction data under different dimensions; wherein the different dimensions at least include: geographical area, enterprise size, enterprise type, transaction platform, and transaction time; Based on the preset networking conditions of different dimensions, establish the network model space corresponding to different dimensions; Projecting the transaction data in different dimensions into the corresponding network model spaces as corresponding nodes to obtain first real-time transaction network models corresponding to different dimensions; Adaptively learn the transaction data under different dimensions to obtain the connection relationship between the nodes in the first real-time transaction network model; the meaning represented by the connection relationship between the nodes in the first real-time transaction network model is associated with the data type of the corresponding dimension.

3. The method for predicting sodium dithionite price trend according to claim 2, characterized in that: The adaptive learning of the transaction data in different dimensions to obtain the connection relationship between the nodes in the first real-time transaction network model includes: Each geographical region in the geographical region trading network is set as a node, and the trade relations between different geographical regions are set as edges; The weight of each edge is set according to the trade volume and transaction frequency of the geographical area; the weight of the edge between the i-th node and the j-th node is w ij It is expressed as: α is an adjustable parameter, which is not less than 0 and not greater than 1. ij is the trade volume between region i corresponding to the i-th node and region j corresponding to the j-th node, F ij is the transaction frequency between region i corresponding to the i-th node and region j corresponding to the j-th node, represents the trade volume V between region i and other regions k ik the sum of is the transaction frequency F between region i and other regions k ik The sum of all regions, k is the sum index, N is the total number of geographic regions, and the value of k ranges from 1 to N.

4. The method for predicting sodium dithionite price trend according to claim 2, characterized in that: The adaptive learning of the transaction data in different dimensions to obtain the connection relationship between the nodes in the first real-time transaction network model includes: Each enterprise in the enterprise transaction network is set as a node, and the transaction relationships between different enterprises are used as edges; Based on the transaction amount, transaction time, and transaction frequency between each enterprise and other enterprises, the weight of the edge corresponding to each enterprise is set; Among them, the weight w of the edge between the mth node and the nth node mn It is expressed as: Among them, β1, β2, and β3 are adjustable parameters, and the sum of β1, β2, and β3 is 1. mn is the transaction amount between enterprise m corresponding to the mth node and enterprise n corresponding to the nth node, T mn F is the countdown from the last transaction time to the current time. mn is the transaction frequency between enterprise m corresponding to the mth node and enterprise n corresponding to the nth node, represents the transaction amount A between enterprise m and other enterprise k mk the sum of represents the inverse of the transaction time T between enterprise m and other enterprises k mk the sum of is the transaction frequency F between enterprise m and other enterprises k mk The sum of all, k is the sum index, M is the total number of enterprises, and the value range of k is from 1 to M.

5. The method for predicting sodium dithionite price trend according to claim 2, characterized in that: The adaptive learning of the transaction data in different dimensions to obtain the connection relationship between the nodes in the first real-time transaction network model includes: Each trading subject in different trading platforms in the platform trading network is set as a node, and the trading relationship between each trading subject in different trading platforms is used as an edge; The weight of each edge is set according to the user flow and product information sharing between various transaction entities; Among them, the weight w of the edge between the pth node and the qth node in the current trading platform is pq It is expressed as: Among them, γ is an adjustable parameter, γ is not less than 0 and not greater than 1, U pq is the user flow between the transaction subject p corresponding to the p-th node and the transaction subject q corresponding to the q-th node, S pq is the degree of commodity information sharing between the transaction subject p corresponding to the p-th node and the transaction subject q corresponding to the q-th node, Represents the user flow U between transaction subject p and other subjects k pk the sum of is the degree of commodity information sharing S between transaction subject p and other subjects k pk The sum of, k is the sum index, L is the total number of entities in the current trading platform, k ranges from 1 to L, λ t is the time decay factor at the current time t, R p and R q They are the platform reputation ratings corresponding to transaction subject p and transaction subject q respectively.

6. The method for predicting sodium dithionite price trend according to claim 1, characterized in that: The change trend prediction model at least includes the following structures: input layer, position encoding layer, adaptive prediction layer, output layer; The change trend prediction model is adopted to analyze the real-time monitoring data stream corresponding to the attribute information in the real-time transaction network model to obtain the change trend characteristics of the sodium dithionite transaction matching the target object, including: Extracting the real-time monitoring data stream corresponding to the attribute information from the real-time transaction network model through the input layer; Through the position coding layer, the position of each node in the real-time monitoring data stream is position-coded to obtain position coding information corresponding to each node; Through the adaptive prediction layer, based on the position coding information, the first attribute feature information corresponding to each node and the second attribute feature information between each node are captured in parallel, and the transaction trend feature corresponding to each node is adaptively predicted based on the captured dependency relationship; Through the output layer, the transaction trend features corresponding to each node are merged into the transaction trend change features.

7. The method for predicting sodium dithionite price trend according to claim 1, characterized in that: After the corresponding sodium dithionite price change trend information is generated based on the transaction trend change feature and displayed to the target object, the method further includes: Identifying abnormal feature points in the transaction trend change characteristics; Based on the abnormal feature points, retrieving corresponding historical transaction data from the real-time transaction network model; Based on the historical transaction data, candidate anomaly types corresponding to the abnormal feature points are determined, and corresponding price warning information is generated, so that the target object can learn the corresponding potential transaction risks based on the price warning information.

8. A sodium dithionite price trend prediction system, characterized in that: The system comprises the following units, wherein: The collection unit is configured to obtain multi-dimensional transaction data of sodium dithionite; the multi-dimensional transaction data at least includes: transaction data of different geographical regions, transaction data of enterprises of different sizes, and transaction data of different transaction platforms; A networking unit is configured to establish a real-time trading network model of sodium dithionite based on the multi-dimensional trading data; the real-time trading network model is provided with a plurality of networking modes, each of which is constructed by using the trading data of a corresponding dimension; the plurality of networking modes at least include: a geographical area trading network, an enterprise trading network, a platform trading network, and a comprehensive trading network; the comprehensive trading network is obtained by using the trading data of at least two dimensions to form a network; The interaction unit is configured to receive a query instruction of a target object to obtain attribute information of the target object; the attribute information includes at least one of the following: a geographical location of the target object, a user type of the target object, an institution to which the target object belongs, and a trading platform available to the target object; The prediction unit is configured to adopt a change trend prediction model to analyze the real-time monitoring data stream corresponding to the attribute information in the real-time transaction network model to obtain the sodium dithionite transaction trend change characteristics matched by the target object; The interaction unit is also configured to generate and display corresponding sodium dithionite price change trend information to the target object based on the transaction trend change characteristics.

9. An electronic device, characterized in that: include: Memory for storing computer software programs; A processor is used to read and execute the computer software program, thereby implementing the sodium dithionite price trend prediction method according to any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, characterized in that: The storage medium stores a computer software program, and when the computer software program is executed by the processor, the method for predicting the price trend of sodium dithionite according to any one of claims 1 to 7 is implemented.

Images