45 results about "Market prediction" patented technology

The invention relates to an intelligent agricultural management system based on a planting-harvesting-distribution full link, and belongs to the field of agriculture, and the system comprises an online seed selection module, an order tracking module, an environment monitoring and early warning module, a pest and disease identification module, a market prediction module, a logistics scheduling module, a product estimation module and a growth state evaluation module. In the planting link, intelligent monitoring and abnormal early warning are carried out on the crop growth environment through the intelligent agent, accurate planting schemes for different crops can be formulated for farmers, and irrigation and fertilization schemes suitable for different regions can be formulated; in the production process, the intelligent agent can intelligently identify bad fruits and give out a treatment scheme, accurately identify diseases and pests and provide an effective prevention and treatment means; in the marketing link, the market trend can be predicted, the production scheme is reasonably arranged, and the production and marketing balance of agricultural products is ensured.

The invention relates to a trading platform that shifts focus from market prediction to disciplined execution, capital management, and trading psychology. By disclosing the general market trend in advance, users are empowered to train and operate within a known directional framework. The system generates synthetic charts using directional random walks anchored at algorithmically or randomly selected convergence points from historical data. This recreates real trends while introducing stochastic variability between anchors. A simulated order book models supply-demand dynamics using Poisson-distributed flows, capturing sentiment, liquidity shifts, and large trades. Features include Accelerated Time Compression for fast-forwarded trade simulation and a Rewind Trade mechanism allowing entry at the start of validated candlesticks. The rewind logic applies to both futures and spot markets, including staking actions. Safeguards such as rewind limits, volume caps, and LP prioritization preserve fairness. The platform supports multi-timeframe execution and bridges strategic training with interactive trading, enhancing decision quality and behavioral discipline.

The invention relates to the technical field of offshore trade platform data analysis methods, and discloses an offshore trade platform data analysis method and system based on a block chain, and the method comprises the steps: constructing a distributed network based on the block chain, encrypting key data in an offshore trade process, uploading the key data to a chain for storage, constructing a data analysis model on the chain, and analyzing the data in the chain. By processing and analyzing the uplink data, decision support such as risk assessment, market prediction and credit rating is provided for the user. According to the method, the security and reliability of the offshore trade data are effectively improved by using the decentralization and tamper-resistant characteristics of the block chain, transparent sharing and efficient collaboration of the data are realized through an intelligent contract and a consensus mechanism, the overall efficiency and competitiveness of the offshore trade are improved, and the healthy development of the international trade is promoted.

A system and method for real-time financial data analysis and market prediction. The system processes diverse inputs, including financial news snippets and trading data, through adaptive codebook generation and codeword allocation. A projection network fuses different data types, creating unified representations for a latent transformer core. The system's architecture enables efficient handling of multi-modal financial data, capturing complex relationships between news sentiment and market behavior. An adaptive codebook generation method ensures the system remains responsive to evolving market conditions. This approach aims to provide more accurate and timely market predictions by leveraging both textual and numerical financial data in a sophisticated, integrated manner.

The invention discloses an agricultural big data multivariate fusion supervision system, and relates to the technical field of agricultural supervision, and the system comprises a soil analysis module which is used for obtaining soil parameter data of an agricultural region, carrying out the analysis of the soil parameter data through a K-Means clustering model, and outputting the quality score and improvement suggestion information of agricultural products in the agricultural region; the meteorological analysis module is used for collecting historical meteorological data of the agricultural region, constructing a time sequence prediction model, predicting a meteorological trend of a future time window and generating meteorological forecast information; through multi-source data fusion of the meteorological analysis module, the agricultural product quality supervision module and the market prediction module, the agricultural product quality, the future meteorological trend and the agricultural product market supply and demand state of a land parcel in an agricultural area are comprehensively judged, and a crop variety recommendation scheme is output based on a decision tree and a scoring function through the planting recommendation module. The decision making is more scientific and accurate, so that the crop yield and the economic benefit are improved.

A system and method for certifying, sharing, and monetizing via marketplaces personal and collective memories as digital assets, integrated with blockchain or other immutable registry technologies. Users upload multimedia memory content, which is authenticated via biometric, contextual, or metadata-based verification. Artificial intelligence assists in structuring the memory and generating immersive presentations incorporating visual elements and optionally one or more sensory modalities including touch, smell, hearing, taste, moisture, nervous system sensations, and temperature. The immersive presentation is then minted as a unique non-fungible token (NFT / nxNFT) or equivalent digital asset, enabling privacy, access control, transferability, and monetization through sale, financing, rental, or co-ownership models. AI further supports market prediction, including dynamic pricing strategies. The system may incorporate a social ecosystem for voting, curation, and emotional engagement, alongside a governance layer for ethical oversight and category designation. Designed for scalability, cross-platform resonance, and anticipates future technologies ensuring interoperability with evolving digital asset standards.

The invention discloses a hierarchical adaptive multi-modal fusion method and system for a multi-modal financial large model, and the method comprises the steps: collecting financial text data, financial image data and time series data, and carrying out the preprocessing; performing modal feature extraction including text features, image features and time sequence features, and generating a text feature vector, an image feature vector and a time sequence feature vector; designing a full connection layer for each modal feature, then connecting a Sigmoid activation function, generating a gating weight vector, and obtaining a weighted feature vector; carrying out information interaction between modes by adopting a multi-head attention mechanism to form a feature vector after interaction enhancement; and constructing a financial domain knowledge graph, generating a domain knowledge vector, and carrying out vector splicing on the feature vector after interaction enhancement and the domain knowledge vector to generate a fusion feature vector. The method aims at improving the understanding and analysis ability of a large language model in application scenes such as financial analysis, risk assessment and market prediction.

The invention provides a carbon market policy impact assessment system based on multi-subject modeling, which relates to the technical field of carbon market prediction and comprises a multi-subject modeling module, a data processing module, a policy scene setting module, a double-layer assessment module, a quantitative index calculation module and a result output module. The multi-subject modeling module is used for constructing an interaction model of a plurality of independent decision subjects in the carbon market; the dynamic interaction model of the independent decision-making main body is constructed through the multi-main-body modeling module, the government, the low-emission enterprises and the high-emission enterprises are all brought into the modeling category, the complex game relation and the policy response mechanism among the main bodies in the carbon market can be effectively revealed, and compared with a traditional macroscopic modeling method, the method has the advantages that the method is simple and convenient to operate. According to the system, the resolution of policy influence factors can be improved, the nonlinear interaction effect of a subject decision is accurately captured, the technical pain point that the policy action process is difficult to restore through traditional evaluation is solved, and core support is provided for deep analysis of policy influence.

The invention discloses a bidding strategy generation method for an electricity-carbon combined market and related equipment, and the method comprises the steps: obtaining real-time market data and historical market data in a preset simulated market environment corresponding to a main intelligent agent and a plurality of other intelligent agents as competition objects, obtaining a power market predicted clearing electricity price and a carbon market predicted price based on historical market data through the prediction model; obtaining subject state information of a subject agent and behavior characteristics of other agents, and obtaining strategy elements based on the obtained data through a trained bidding strategy model; and obtaining a power market quotation strategy and a carbon market transaction strategy based on the strategy elements, and forming a bidding strategy of the power-carbon combined market. According to the method, the bidding strategy of the power generation manufacturer can be gradually optimized by integrating the market data and the competitive behavior characteristics, so that the income capability and the market competitiveness in the electricity-carbon combined market are improved.

The present invention discloses a method and device for optimizing the integrated procurement supply chain based on the industrial internet. The method comprises the following steps: predicting the market price of raw materials; establishing a multi-period dynamic procurement model with the minimum unit raw material cost, taking into account the predicted market price of raw materials, and optimizing and solving the optimal raw material procurement quantity for each procurement cycle; obtaining the historical supply data of each supplier, constructing a supplier automatic evaluation model based on stacked autoencoders and bagging integrated learning, and obtaining the scores of each supplier; comprehensively considering the optimal raw material procurement quantity and supplier scores, establishing a multi-objective optimization model that minimizes procurement costs and maximizes the comprehensive effects of suppliers, and solving and obtaining the optimal procurement strategy from each supplier. The present invention comprehensively acquires, centrally processes and analyzes the entire procurement supply chain process data through the industrial internet platform, comprehensively considers the correlation between different businesses, and can effectively reduce procurement costs and improve procurement efficiency.

The invention relates to the technical field of electricity market prediction, in particular to an electricity market day-ahead electricity price prediction method and system based on an enhanced linear Stacking framework, and the method specifically comprises the steps: obtaining multi-source heterogeneous data, carrying out the preprocessing of the data, obtaining a standardized feature matrix, calculating the feature importance through a random forest, and carrying out the calculation of the feature importance; k-fold time sequence cross validation and marginal benefit analysis are combined to screen an optimal feature subset, two basic learners are constructed, fold time sequence cross validation is adopted to train the basic learners, fold-out prediction vectors are obtained, statistical features are extracted, and the fold-out prediction vectors and the statistical features serve as input to train the basic learners, so that the optimal feature set is obtained. And then the optimal hyper-parameter of the basic learner is searched through Bayesian optimization and re-fitting is completed, the meta learner is trained to obtain a Stacking integrated model, a test sample is input into the model, and a day-ahead electricity price prediction value of an original dimension is obtained through inverse transformation. The method can improve the accuracy and stability of electricity price prediction, and is suitable for an electricity market day-ahead electricity price prediction scene.

The invention discloses a micro-grid system energy storage capacity optimal configuration method, system and device adapting to a multi-element power market, and relates to the technical field of power systems, the method comprises the following steps: acquiring and processing an unbalanced power signal in a micro-grid system, and obtaining a corresponding power instruction according to an energy storage characteristic; and based on the power instruction and the optimal capacity planning configuration, solving a dynamic optimization configuration model, and obtaining a multi-power market day-ahead optimal joint bidding strategy of the micro-grid system. According to the method, the day-ahead optimal joint bidding strategy is obtained based on the power instruction and the optimal capacity planning configuration according to the external multi-power market prediction information participated by the micro-grid system, the internal power balance demand of the micro-grid energy storage system is preferably met, the capacity and power configuration of the micro-grid energy storage system is reasonably configured, and the optimal capacity of the micro-grid energy storage system is optimized. Stable operation of the micro-grid energy storage system is guaranteed, and efficient utilization of energy storage resources of the micro-grid energy storage system is achieved.

The embodiment of the invention provides a resource consumption abnormity early warning method and device, electronic equipment and a storage medium, belongs to the technical field of data processing, and is suitable for the fields of financial science and technology and medical treatment. The method comprises the following steps: constructing a target resource consumption prediction model based on historical resource consumption data and resource market data; performing preliminary resource consumption prediction based on the target resource consumption prediction model and the current resource market data to obtain resource consumption prediction data; acquiring resource consumption real-time data; performing associated data fluctuation prediction based on the historical resource market data and the current resource market data to obtain resource market prediction parameters; performing resource consumption adjustment on the resource consumption prediction data based on the resource market prediction parameters to obtain a resource consumption early warning threshold; and performing resource consumption data abnormity early warning based on the resource consumption early warning threshold and the resource consumption real-time data. According to the embodiment of the invention, the accuracy of the resource consumption abnormity early warning threshold can be improved.

The invention relates to the technical field of electric power market prediction, and discloses an electric power spot transaction node electricity price prediction method, which comprises the following steps of 1, obtaining historical transaction data of an electric power market, operation data of a generator set and interference data influencing electric power market transaction, and classifying the data to form a data set; 2, analyzing the fluctuation condition of the daily electricity price of the electricity market, and generating a corresponding floating index; 3, analyzing the daily supply and demand balance condition of the electricity market and the daily operation fluctuation condition of the generator set, and generating a corresponding monitoring data set; 4, setting a time axis, drawing a curve graph, and evaluating the relevance between different parameters and the electricity price; 5, setting a floating threshold value, judging a time node of abnormal fluctuation of the electricity price, and marking the time node in the curve graph; and step 6, constructing a prediction model according to the curve graph, the market data set, the monitoring data set and the influence data set.

The invention belongs to the field of financial science and technology and quantitative investment, and relates to a stock market prediction method based on a hidden Markov model and dynamic similarity measurement. The method comprises the following steps: firstly, modeling by utilizing trend information of HMM historical stock price data, and obtaining a hidden state sequence through state mapping; secondly, calculating the increase and decrease amplitude of the closing price, constructing increase and decrease information, and performing feature fusion on the increase and decrease amplitude and hidden state vectors generated by the model in the corresponding time step length so as to construct feature expression capable of reflecting market dynamics and representing potential modes; then, training is carried out under different random seeds by utilizing an HMM model, and the seeds with relatively stable effects are selected as fixed seeds for subsequent model training; and then, with maximization of prediction accuracy as a target, selecting an optimal distance by cyclically traversing a distance threshold. And according to the obtained optimal distance, screening out neighbors with relatively high credibility for the target sample, and naming the neighbors as credible neighbors. And finally, weighting and aggregating the labels of the credible neighbors to obtain a final prediction result.

The invention relates to the technical field of bulk commodity operation, in particular to a data sharing and analysis method and system based on multi-dimensional association, and the method comprises the steps: collecting multi-source heterogeneous data through a mixed mode of interface synchronization and manual reporting, and quantifying an unstructured text into a standardized index through a bidirectional LSTM semantic analysis algorithm; a cosine similarity and grey relational degree fusion algorithm and an LSTM dynamic weight optimization model are adopted, association rules are mined in combination with an improved Apriori algorithm, and market pre-judgment is achieved through three sub-models of supply and demand, price and risk; and finally, establishing a closed-loop mechanism, and outputting an inventory adjustment scheme and a futures hedging strategy. The system correspondingly comprises four functional modules, full-process data processing and decision support are achieved, and scientificity and operability of operation decision are improved.

The invention discloses an electricity transaction agent market prediction method and system based on graph deep learning, and relates to the technical field of electricity transaction, and the method comprises the steps: constructing an electricity market transaction graph comprising a node set and an edge set; performing node and structural feature coding on the electricity market transaction graph, and extracting global embedded representation of the market; and modeling market subjects corresponding to the nodes into power transaction agents, and executing multi-subject strategy interaction learning among the power transaction agents by using global embedded representation and a historical strategy sequence of the power transaction agents. Balance evolution is carried out in the graph reinforcement learning process to dynamically update the strategy weight and the market response threshold value of the power transaction agent; and performing behavior simulation deduction by using the updated power transaction agent to establish a time sequence prediction result. The technical problem that the market price fluctuation prediction precision is insufficient due to the fact that the electricity market transaction relation is complex and the subject behavior is difficult to accurately model in the prior art is solved, and the technical effect of improving the market prediction precision is achieved.

The invention relates to the technical field of solid waste resourceful treatment, and discloses a multi-property discarded equipment disassembly product ownership binding and transaction collaboration system, which comprises an online sensing and real-time ownership generation module used for sensing physical attributes of products in real time to judge grades and generating atomic ownership records associated with source asset identifiers; the predictive dynamic packaging and ownership solidification module decides independent or mixed packaging according to a dynamic packaging threshold value, and calculates and solidifies the ownership proportion of each property right party during mixing; and the aggregated transaction and intelligent contract automatic clearing module is used for performing transaction on aggregated products, a built-in machine learning strategy optimization engine of the aggregated transaction and intelligent contract automatic clearing module performs market prediction based on internal and external data, generates the dynamic packaging threshold value and feeds back the dynamic packaging threshold value to the packaging module, and closed-loop control is formed. According to the method, dynamic ownership calculation and aggregation transaction intelligent contracts are utilized, under the premise of no physical isolation, ownership is accurate, and the overall value of multi-party asset disposal is improved.

This invention belongs to the field of financial investment decision-making technology, specifically a financial investment decision-making assistance method and system based on game theory and multi-source big data. It collects and processes multi-source financial data, constructs a unified data foundation, and uses machine learning to generate multi-dimensional market prediction signals based on this foundation. These prediction signals are then input into a game theory model to simulate the behavior of market participants, solve for equilibrium strategies, and generate investment or risk control suggestions based on the game results, which are then visualized. This invention overcomes the limitations of single data sources by integrating multi-source big data and generating multimodal prediction signals. By combining game theory to model the complex interactive behaviors of market participants, it can deduce market evolution trends from a global perspective, thereby assisting in the formulation of more forward-looking investment or risk control strategies. It can simulate information asymmetry, strategy interaction, and dynamic evolution processes in the market, thus providing a better understanding of market fluctuations and the impact of sudden events.

The application discloses a social media-based interpretable dynamic graph network benefit prediction model training method, relates to the technical field of text analysis and market prediction, and obtains market data and related social media data of a financial asset to divide the data by weeks; for the social media data of each week, a topic is taken as a node, and similarity between topics is taken as an edge weight, a topic correlation graph of each week is constructed to obtain weekly graph data; a prediction model performs time sequence updating and market prediction according to the weekly graph data; a text time sequence memory unit obtains a current week topic memory vector according to a current week topic text embedding vector and in combination with a last week topic memory vector; a multi-layer graph attention network captures the correlation features between nodes through an attention mechanism and adopts attention weights for weighted summation to obtain a current week global graph embedding; and a classifier outputs a next week price change prediction result. The application solves the problem that the prior art cannot effectively capture the time sequence evolution of topic sentiment and the correlation between topics.

The invention discloses a carbon transaction market dynamic price prediction method and system based on artificial intelligence, and relates to the technical field of market prediction, and the method comprises the steps: collecting carbon transaction event data, calculating the length of a dynamic event window according to the price fluctuation and the event frequency, analyzing the price change rate through a first-order difference, and determining a window starting point; forming unequal-length windows, and for different windows, calculating importance scores according to the price change speed of the transaction event and the transaction volume. According to the method, the rapid identification effect of important market change points is achieved through combination of first-order differential price change rate calculation and dynamic time window adjustment, the effect of improving the reconstruction quality of key information of transaction events is achieved through combination of a resampling set and window importance weight calculation, and the method is suitable for popularization and application. Through combining long-term feature extraction of a comprehensive price mean value and a price change rate with calculation of short-term disturbance features, an effect of simultaneously extracting core dynamic features from local market behaviors and an overall market trend is achieved.

The invention discloses a power market supply and demand prediction system and method based on multi-source data fusion and an LSTM neural network, and relates to the technical field of power market prediction, and the method comprises the steps: collecting multi-source data, carrying out the time alignment and formatting, and generating standardized time series data; carrying out missing value filling, anomaly detection and denoising processing on the standardized time sequence data to generate a time sequence feature set; time sequence features of different time scales are extracted from the high-quality fusion feature sequence, a semantic memory LSTM network is constructed, output of each scale is fused through attention gating, and an initial supply and demand prediction result is generated; and calculating a residual error according to the difference between the initial supply and demand prediction result and the actual observation value, when the residual error exceeds a residual error triggering threshold value, triggering the correction network to generate a correction item, and superposing the correction item with the initial prediction result to generate a final supply and demand prediction result. According to the invention, power market scheduling, load management and new energy consumption optimization are facilitated.

The application belongs to the technical field of electric power data analysis, and discloses a power market energy storage data interaction feature analysis method and system, which comprises the following steps: collecting original data of a power market, energy storage equipment and environmental conditions, identifying and cleaning abnormal data in the original data through data preprocessing to generate a data set in a unified format; extracting relevant features of each type of original data in the data set, eliminating invalid features in the relevant features based on variance threshold analysis, and retaining effective features to form a feature set; based on a time series prediction algorithm, a power market prediction model is constructed to predict future demand data and power price fluctuations of the power market and evaluate market risks existing in the energy storage system. The application combines data collection, preprocessing, feature extraction, market prediction, optimization decision and dynamic feedback mechanism, and can effectively improve the scheduling accuracy and efficiency of the energy storage system.

The invention relates to the technical field of business intelligence and prediction analysis, in particular to a cigarette operator confidence index analysis and decision prediction method. The method comprises the following steps: acquiring user data, preprocessing the user data and calculating a standardized influence degree; performing dynamic aggregation to obtain first-level indexes, performing relation analysis verification to obtain prediction indexes, and dividing the prediction indexes to obtain confidence prediction state comprehensive indexes and market prediction state comprehensive indexes; calculating to obtain a demand prediction index, and then performing state division to obtain a demand prediction state index; and generating a three-dimensional decision prediction matrix based on the three state indexes so as to obtain decision prediction of the cigarette operator. In this way, dynamic association monitoring, accurate association verification, trend association prediction and intelligent association decision support of the confidence index of the cigarette operator can be achieved, and scientificity and foresight of operation decision are effectively improved.

The invention discloses an enterprise customer information management system and management method, and relates to the technical field of enterprise information management. The system comprises a customer basic information module, a space-time sequence analysis module, a geographic information association module, a data storage module and an interactive display module, the customer basic information module accurately processes missing values and abnormal values, and data quality and management efficiency are improved in combination with distributed storage; the LSTM model integrates multi-dimensional features, monthly updating ensures accurate purchase time prediction, and assists an enterprise in reasonable planning; the geographic information association module visually presents the relationship between the regional value and the associated enterprise through coordinate transformation, clustering analysis and GIS visualization, and optimizes resource configuration; in customer management, the system outputs purchase prediction and triggers collaborative early warning, and customer interaction is enhanced; a data verification feedback mechanism forms a closed loop, and the model is continuously optimized; the modules are flexible in communication and friendly in interface, and customer control, market prediction and decision-making capabilities of enterprises are comprehensively improved.

The embodiment of the invention provides a laser market prediction information generation method and device. The method comprises the following steps: acquiring multi-dimensional market information corresponding to a laser in response to a received market prediction request for the laser; preprocessing the multi-dimensional market information corresponding to the laser, and classifying the preprocessed multi-dimensional market information according to a time dimension, a product type dimension and a region dimension to obtain multiple types of text information; performing feature extraction on the plurality of types of text information to obtain index information corresponding to the plurality of types of text information, and constructing knowledge graph information corresponding to the laser based on the index information corresponding to the plurality of types of text information; and calling the trained market prediction model according to the knowledge graph information corresponding to the laser, and generating market prediction information corresponding to the laser. The method improves the accuracy of market prediction information.

This invention provides a method, system, device, and storage medium for power system risk control decision-making, belonging to the field of power system technology. The method includes: acquiring power system fusion data and power market text information, and labeling the text information with market sentiment tags; training an optimal electricity price sentiment mapping and an optimal sentiment mapping correction quantity based on the tags and text information, and extracting market sentiment features based on the optimal mapping and correction quantity; acquiring a market feedback time decay mapping based on the aforementioned text information, and fusing the text information, time decay mapping, and market sentiment features to obtain power market expectation features; then combining the power fusion data to predict the base predicted electricity price and the market predicted electricity price; and obtaining the base price tendency quantity under gating parameters based on the predicted electricity price to obtain the estimated electricity price; and generating a power anomaly alarm based on the estimated electricity price for risk control decision-making. This invention can improve the reliability of power system risk control decision-making.

The application discloses an enhanced model day-ahead electricity price prediction method based on dynamic feature selection, which extracts features from the original data set through a multi-head attention mechanism for multidimensional correlation analysis, generates a feature weight vector in real time to realize dynamic screening of the optimal feature subset, thereby suppressing redundant information and nonlinear noise features at the source; then, an enhanced XGBoost model is used to deeply mine the structured relationship of power market data, simultaneously combined with a Transformer model to capture the long-range time series dependence characteristics of the electricity price sequence, and through a gating mechanism to perform dynamic weighted fusion, to realize the complementary integration of spatial features and time series features; finally, a multi-objective Bayesian optimization algorithm is introduced to obtain a trained enhanced model through multi-objective collaborative optimization, and then prediction is performed. The application significantly improves the prediction accuracy and robustness, effectively solves the deployment problem of the model in a resource-limited environment, and has high engineering application value and power market prediction practicality.

Limited data availability poses a major obstacle in training deep learning models for financial applications. Synthesizing financial time series (FTS) to augment real-world data is challenging due to irregular and scale-invariant patterns associated with FTS—temporal dynamics that repeat with varying duration and magnitude. A novel generative framework called FTS-Diffusion is developed, consisting of three modules to model irregular and scale-invariant patterns. First, a scale-invariant pattern recognition algorithm extracts recurring patterns that vary in duration and magnitude. Second, a pattern-conditioned diffusion network synthesizes segments of patterns. Third, the temporal evolution of patterns is modeled in order to aggregate the generated segments. Extensive experiments show that FTS-Diffusion generates synthetic FTS highly resembling observed data, outperforming state-of-the-art alternatives. Two downstream experiments demonstrate that augmenting real-world data with synthetic data generated by FTS-Diffusion reduces the error of stock market prediction by up to 17.9%.

The application discloses an enterprise customer information management system and a management method, and relates to the technical field of enterprise information management; the system comprises a customer basic information module, a space-time sequence analysis module, a geographic information correlation module, a data storage module and an interactive display module; the customer basic information module accurately processes missing values and abnormal values, and improves data quality and management efficiency in combination with distributed storage; an LSTM model integrates multi-dimensional features, monthly updates guarantee procurement time prediction accuracy, and helps enterprises to reasonably plan; the geographic information correlation module directly presents regional value and the relationship between related enterprises through coordinate conversion, clustering analysis and GIS visualization, and optimizes resource allocation; in customer management, the system outputs procurement prediction and triggers collaborative early warning, and enhances customer interaction; a data verification feedback mechanism forms a closed loop, and continuously optimizes the model; each module flexibly communicates, the interface is friendly, and the enterprise customer control, market prediction and decision-making ability are comprehensively improved.