Logistics demand forecasting method and related equipment based on big data analysis

Abstract

The present invention relates to a logistics demand forecasting method based on big data analysis, comprising the following steps: using a preset principal component analysis algorithm to extract logistics features from the multi-source collected data to obtain logistics extraction features; performing multi-dimensional analysis on the logistics extraction features to obtain multi-dimensional analysis results; identifying logistics demand based on the multi-dimensional analysis results and the logistics extraction features to obtain fluctuations in logistics demand, and generating visualization charts based on the fluctuations; drawing a logistics demand trend curve and a logistics demand fluctuation pattern curve based on the visualization chart; using a preset prediction algorithm to predict logistics types based on the logistics demand trend curve and the logistics demand fluctuation pattern curve to obtain the type of logistics demand, and formulating corresponding logistics demand strategies based on the type of logistics demand; and solving the technical problems of inaccurate logistics demand forecasting and unreasonable resource allocation.

Description

Technical Field

[0001] The present invention relates to the field of logistics demand technology, and in particular to a logistics demand forecasting method based on big data analysis and related equipment. Background Art

[0002] In traditional logistics management, for complex supply chain networks, logistics needs are usually determined by manual planning and operator experience. This approach is not only inefficient, but also difficult to ensure the consistency and accuracy of logistics activities when dealing with logistics activities with high precision requirements. In addition, this manual method of planning logistics needs may also lead to delays or other operational problems in the logistics process.

[0003] As the logistics industry develops towards automation and intelligence, the demand for logistics accuracy and efficiency is increasing. Traditional logistics management methods can no longer meet the needs of the modern logistics industry. Although existing automated logistics technology can improve logistics efficiency to a certain extent, there are still problems such as inaccurate demand forecasting and unreasonable resource allocation. Especially for market demands with high uncertainty, existing technologies often require a lot of pre-processing work, such as manual forecasting of demand, which not only increases the preparation time before logistics activities, but also may introduce additional human errors. Summary of the invention

[0004] The main purpose of the present invention is to provide a logistics demand forecasting method and related equipment based on big data analysis, which solves the technical problems of inaccurate logistics demand forecasting and unreasonable resource allocation.

[0005] To achieve the above object, the present invention provides a logistics demand forecasting method based on big data analysis, comprising the following steps:

[0006] Collect multiple data sources in the logistics supply chain to obtain multi-source collected data; wherein the multi-source collected data includes order information, inventory levels and transportation records;

[0007] Using a preset principal component analysis algorithm to extract logistics features from the multi-source collected data, to obtain logistics extraction features; wherein the logistics extraction features include demand fluctuations and time series trends;

[0008] Performing a multi-dimensional analysis on the logistics extraction features to obtain a multi-dimensional analysis result; wherein the multi-dimensional analysis result includes demand fluctuation frequency, seasonal trend analysis, regional demand difference analysis, product category demand changes, customer demand behavior patterns and transportation mode efficiency comparison;

[0009] Based on the multi-dimensional analysis results and the logistics extraction features, the logistics demand is identified to obtain the fluctuation of the logistics demand, and a visual chart is generated based on the fluctuation;

[0010] Draw a logistics demand trend curve and a logistics demand fluctuation pattern curve based on the visualization chart;

[0011] By using a preset prediction algorithm, a logistics type prediction is performed based on the logistics demand trend curve and the logistics demand fluctuation pattern curve to obtain the type of logistics demand, and a corresponding logistics demand strategy is formulated based on the type of logistics demand.

[0012] Furthermore, multiple data sources in the logistics supply chain are collected to obtain multi-source collected data, including:

[0013] Collect multiple data sources in the logistics supply chain to obtain collected data;

[0014] Performing standardization processing on the collected data to obtain standardized data;

[0015] Identifying and removing duplicates from the standardized data to obtain identified and removed duplicates data;

[0016] The identified and deduplicated data are classified to obtain multi-source collected data.

[0017] Furthermore, a preset principal component analysis algorithm is used to extract logistics features from the multi-source collected data to obtain logistics extraction features, including:

[0018] Performing logistics feature extraction on the multi-source collected data to obtain preliminary extracted features, and performing vector conversion on the preliminary extracted features to obtain a multi-source collected vector;

[0019] Inputting the multi-source acquisition vector into a preset principal component analysis algorithm to perform principal component analysis to obtain a principal component analysis result;

[0020] Constructing a feature matrix based on the principal component analysis result, and calculating a covariance matrix based on the feature matrix; wherein the covariance matrix is ​​used to describe the linear relationship between each logistics extraction feature and the strength of the logistics extraction feature;

[0021] Performing eigenvalue decomposition on the covariance matrix to obtain principal component eigenvectors and eigenvalues ​​corresponding to the principal component eigenvectors;

[0022] Calculating the variance explanation ratio of the principal component analysis result based on the principal component eigenvector and the eigenvalue corresponding to the principal component eigenvector to evaluate the importance of each principal component analysis result in the multi-source collected data and obtain the variance explanation power of each principal component analysis result;

[0023] By using a preset load weight extraction algorithm, based on each of the variance explanatory power, load weight calculation is performed on the preliminary extracted features to obtain a weight coefficient of each preliminary extracted feature;

[0024] Setting a weight coefficient threshold, removing the weight coefficients of each initially extracted feature that is lower than the weight coefficient threshold, and obtaining a standard weight coefficient;

[0025] The standard weight coefficient is matched with the logistics extraction feature to obtain the corresponding logistics extraction feature.

[0026] Furthermore, the logistics extraction features are subjected to multi-dimensional analysis to obtain multi-dimensional analysis results, including:

[0027] Performing a multi-dimensional analysis on the logistics extraction features through a preset hidden Markov model to obtain a preliminary multi-dimensional analysis result;

[0028] Based on the multi-dimensional analysis results, the hidden patterns of logistics demand are identified to reveal the driving factors of changes in logistics demand;

[0029] Using a preset Bayesian network, conditional dependency inference is performed on the logistics extraction features based on the driving factors to obtain an inference result;

[0030] Using a preset interactive visualization technology, the correlation lines between the inference result and the preliminary multi-dimensional analysis result are displayed to obtain a conditional dependency relationship diagram;

[0031] The conditional dependency diagram is interpreted to obtain multi-dimensional analysis results.

[0032] Furthermore, based on the multi-dimensional analysis results and the logistics extraction features, logistics demand identification is performed to obtain the fluctuation of logistics demand, and a visualization chart is generated based on the fluctuation, including:

[0033] Based on the multi-dimensional analysis results, the logistics demand is identified to obtain logistics demand information; wherein the logistics demand information includes demand quantity, demand frequency, demand cycle and demand urgency;

[0034] quantifying the logistics demand information to obtain the fluctuation range of the logistics demand information;

[0035] Using the preset data genomics technology to perform pattern recognition on the logistics extraction features, the fundamental characteristics of logistics demand are obtained;

[0036] Based on the fundamental characteristics and the fluctuation range, the logistics demand is deeply analyzed to obtain the fluctuation of the logistics demand;

[0037] Determine whether there are abnormal fluctuations in the fluctuation situation. If so, visualize the fluctuation situation of the logistics demand based on the abnormal fluctuations to obtain a dynamic fluctuation curve graph, and use the dynamic fluctuation curve graph as a visualization chart; wherein, the abnormal fluctuations include a surge in demand, a sudden drop in demand and abnormal periodic fluctuations.

[0038] Furthermore, drawing a logistics demand trend curve and a logistics demand fluctuation pattern curve based on the visualization chart includes:

[0039] Deconstructing the data in the visualization chart to obtain deconstructed data;

[0040] Identifying long-term and / or short-term trends from the deconstructed data; using the long-term and / or short-term trends as logistics demand trends;

[0041] Draw a logistics demand trend curve based on the logistics demand trend;

[0042] The visualization chart is interpreted by Fourier transform to obtain the frequency components of the demand changes; wherein the frequency components include fundamental frequency and harmonics;

[0043] Analyzing the amplitudes of the different frequency components to obtain amplitude analysis results;

[0044] Constructing a frequency spectrum based on the amplitude analysis result, and obtaining a fluctuation pattern of each frequency component to logistics demand based on the frequency spectrum;

[0045] A logistics demand fluctuation pattern curve is drawn based on the logistics demand fluctuation pattern.

[0046] Furthermore, the logistics demand trend and the logistics demand fluctuation pattern are predicted by a preset prediction algorithm to obtain the type of logistics demand, and a corresponding logistics demand strategy is formulated based on the type of logistics demand, including:

[0047] The logistics demand trend is predicted by a preset prediction algorithm to obtain the dynamic demand of logistics demand; wherein the dynamic demand includes seasonal fluctuations, market penetration rate and supply chain dynamics;

[0048] Predicting the logistics demand fluctuation pattern to obtain a demand change pattern; wherein the demand change pattern includes a stable pattern, a growth pattern, a fluctuation pattern and a sudden pattern;

[0049] Based on the dynamic demand of the logistics demand and the demand change pattern, the type of the logistics demand is predicted to obtain the type of the logistics demand;

[0050] Formulate a preliminary logistics strategy based on the type of logistics demand; wherein the preliminary logistics strategy includes inventory management, transportation planning and resource allocation;

[0051] Conducting risk assessment on the implementation process of the preliminary logistics strategy to obtain risk assessment results;

[0052] The preliminary logistics strategy is optimized based on the risk assessment result to obtain a corresponding logistics demand strategy.

[0053] The present invention also provides a logistics demand prediction device based on big data analysis, comprising:

[0054] A collection module is used to collect multiple data sources in the logistics supply chain to obtain multi-source collected data; wherein the multi-source collected data includes order information, inventory levels and transportation records;

[0055] An extraction module, used to extract logistics features from the multi-source collected data using a preset principal component analysis algorithm to obtain logistics extraction features; wherein the logistics extraction features include demand fluctuations and time series trends;

[0056] An analysis module, used to perform multi-dimensional analysis on the logistics extraction features to obtain multi-dimensional analysis results; wherein the multi-dimensional analysis results include demand fluctuation frequency, seasonal trend analysis, regional demand difference analysis, product category demand changes, customer demand behavior patterns and transportation mode efficiency comparison;

[0057] An identification module, used to identify logistics demand based on the multi-dimensional analysis results and logistics extraction features, obtain the fluctuation of logistics demand, and generate a visual chart based on the fluctuation;

[0058] A drawing module, used for drawing a logistics demand trend curve and a logistics demand fluctuation pattern curve based on the visualization chart;

[0059] The prediction module is used to predict the logistics type based on the logistics demand trend curve and the logistics demand fluctuation pattern curve through a preset prediction algorithm, obtain the type of logistics demand, and formulate a corresponding logistics demand strategy based on the type of logistics demand.

[0060] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

[0061] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned methods are implemented.

[0062] The logistics demand forecasting method based on big data analysis provided by the present invention comprises the following steps: collecting multiple data sources in the logistics supply chain to obtain multi-source collected data; using a preset principal component analysis algorithm to extract logistics features from the multi-source collected data to obtain logistics extraction features; performing multi-dimensional analysis on the logistics extraction features to obtain multi-dimensional analysis results; identifying logistics demand based on the multi-dimensional analysis results and the logistics extraction features to obtain the fluctuation of logistics demand, and generating a visualization chart based on the fluctuation; drawing a logistics demand trend curve and a logistics demand fluctuation pattern curve based on the visualization chart; using a preset prediction algorithm, predicting the logistics type based on the logistics demand trend curve and the logistics demand fluctuation pattern curve to obtain the type of logistics demand, and formulating a corresponding logistics demand strategy based on the type of logistics demand. Through the above technical solution, the technical problems of inaccurate logistics demand forecasting and unreasonable resource allocation are solved, the accuracy and efficiency of logistics demand forecasting are improved, and human intervention is reduced, thereby optimizing the configuration and utilization of logistics resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 It is a schematic diagram of the steps of a logistics demand prediction method based on big data analysis in one embodiment of the present invention;

[0064] Figure 2 It is a structural block diagram of a logistics demand prediction device based on big data analysis in one embodiment of the present invention;

[0065] Figure 3 It is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

[0066] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0067] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0068] like Figure 1 As shown, Figure 1 It is a schematic diagram of the steps of a logistics demand prediction method based on big data analysis in one embodiment of the present invention;

[0069] In one embodiment of the present invention, a logistics demand forecasting method based on big data analysis is provided, comprising the following steps:

[0070] Step S1, collecting multiple data sources in the logistics supply chain to obtain multi-source collected data; wherein the multi-source collected data includes order information, inventory levels and transportation records.

[0071] Specifically, multiple data sources in the logistics supply chain are collected: this means collecting relevant data from different links and different participants in the logistics supply chain. These data sources include but are not limited to point-of-sale systems, warehouse management systems, transportation management systems, customer relationship management systems, etc. Obtain multi-source collected data: Through the above collection process, a comprehensive data set will be obtained, which contains information from all aspects of the supply chain. The multi-source collected data includes order information: order information refers to the specific details of the customer's order, including but not limited to order number, type of ordered product, quantity, order time, expected delivery date, etc. The multi-source collected data includes inventory level: inventory level refers to the inventory of various products in the warehouse at a given point in time. These data can help understand the current inventory status and whether there is enough inventory to meet future order needs. The multi-source collected data includes transportation records: transportation records cover all data related to the transportation of goods, such as transportation mode (such as road, rail, sea or air), transportation cost, transportation time, carrier information, cargo status (such as in transit, delivered, etc.), etc. Through the above steps, we can obtain comprehensive and detailed logistics supply chain data, which will become the basis for subsequent analysis, helping us better understand the changing patterns of logistics demand and make more accurate predictions based on this.

[0072] Step S2, using a preset principal component analysis algorithm to extract logistics features from the multi-source collected data to obtain logistics extraction features; wherein the logistics extraction features include demand fluctuations and time series trends.

[0073] Specifically, using a preset principal component analysis algorithm: Here, it refers to using a preset principal component analysis (PCA) algorithm. PCA is a common dimensionality reduction technique used to extract the most important features from the original data while reducing the dimensions of the data. Performing logistics feature extraction on the multi-source collected data: This means inputting the previously collected multi-source collected data into the PCA algorithm, with the purpose of extracting the features most relevant to logistics demand forecasting from these data. Obtaining logistics extraction features: Through the application of the PCA algorithm, we can extract a set of new features from the original multi-source collected data, which can better reflect the key information of logistics demand. The logistics extraction features include demand fluctuations: Demand fluctuations refer to the changes in logistics demand over time, usually manifested as an increase or decrease in demand. One of the features extracted by the PCA algorithm is demand fluctuations, which helps us understand the law of changes in logistics demand over time. The logistics extraction features include time series trends: Time series trends refer to a long-term trend of changes in logistics demand over time. By analyzing the logistics extraction features, the overall trend of logistics demand over time can be identified, which is crucial for predicting future logistics demand. Through the above steps, the principal component analysis algorithm is used to extract features closely related to logistics demand forecasting from multi-source collected data. These features will be used in subsequent analysis and forecasting processes to help us better understand and predict changes in logistics demand.

[0074] Step S3, performing a multi-dimensional analysis on the logistics extraction features to obtain a multi-dimensional analysis result; wherein the multi-dimensional analysis result includes demand fluctuation frequency, seasonal trend analysis, regional demand difference analysis, product category demand changes, customer demand behavior patterns and transportation mode efficiency comparison.

[0075] Specifically, the logistics extraction features are subjected to multi-dimensional analysis: this step means using statistical analysis and other data mining techniques to conduct in-depth analysis on the previously extracted logistics features to reveal the logistics demand characteristics in different dimensions. Obtain multi-dimensional analysis results: through multi-dimensional analysis, we can obtain a series of analysis results on the characteristics of logistics demand, which will help us better understand the changing laws of logistics demand. The multi-dimensional analysis results include demand fluctuation frequency: demand fluctuation frequency refers to the number and intensity of fluctuations in logistics demand within a certain time range. By analyzing the demand fluctuation frequency, we can understand the stability of logistics demand and possible cyclical changes. The multi-dimensional analysis results include seasonal trend analysis: seasonal trend analysis refers to the law of seasonal changes in logistics demand. By analyzing the changes in logistics demand in different seasons, we can predict the peak or trough of demand in a specific season. The multi-dimensional analysis results include regional demand difference analysis: regional demand difference analysis refers to the changes in logistics demand between different geographical regions. By analyzing the differences in logistics demand in different regions, we can optimize the regional allocation of logistics resources. The multi-dimensional analysis results include product category demand changes: product category demand changes refer to the changes in demand for different product categories over time. By analyzing the demand for different types of products, we can better plan inventory and supply chain management. The multi-dimensional analysis results include customer demand behavior patterns: customer demand behavior patterns refer to the changes in customer purchasing behavior over time. By analyzing customer demand behavior, we can predict future demand trends and optimize customer service. The multi-dimensional analysis results include transportation mode efficiency comparison: transportation mode efficiency comparison refers to the comparative analysis of different transportation modes (such as sea transportation, air transportation, land transportation, etc.) in terms of cost, speed, reliability, etc. By analyzing the efficiency of transportation modes, we can choose the most suitable transportation mode to improve overall logistics efficiency. Through the above steps, we can deeply understand the characteristics of logistics demand from multiple perspectives, and these analysis results will provide strong support for us to formulate more accurate logistics demand forecasts and strategies.

[0076] Step S4, identifying logistics demand based on the multi-dimensional analysis results and logistics extraction features, obtaining fluctuations in logistics demand, and generating a visualization chart based on the fluctuations.

[0077] Specifically, based on the multi-dimensional analysis results and logistics extraction features, logistics demand identification is performed: In this stage, all multi-dimensional analysis results obtained in the previous step and the previously extracted logistics features will be used to identify logistics demand. This means that the information of all analysis dimensions will be comprehensively considered to determine the specific manifestations of logistics demand and its changing rules. Obtain the fluctuation of logistics demand: Through comprehensive analysis, it will be possible to determine the specific fluctuations of logistics demand at different time points, in different regions, and in different product categories. These fluctuations can help understand the change pattern of logistics demand over time, such as the time distribution of peaks and troughs, periodic changes, etc. Generate visual charts based on the fluctuations: In order to more intuitively display the fluctuations of logistics demand, the analysis results will be presented in the form of charts. These charts include but are not limited to time series charts, bar charts, heat maps, etc., which can help users quickly understand the dynamic changes of logistics demand and provide support for decision-making. Specifically, we can implement this step in the following aspects: Integrate multi-dimensional analysis results: First, all relevant multi-dimensional analysis results need to be integrated to ensure that the data of each analysis dimension is ready. Apply statistical methods to identify demand: Statistical methods and machine learning algorithms are used to identify patterns and trends in logistics demand. For example, time series analysis can be used to capture how demand changes over time. Analyze fluctuations in logistics demand: Based on the analysis results, clarify the fluctuation characteristics of logistics demand, such as the time points when the peak and valley values ​​of demand occur, the amplitude of fluctuations, etc. Design and generate visual charts: Design appropriate chart types based on the fluctuations, and use corresponding software tools (such as Excel, Tableau or other data visualization tools) to generate these charts. Review and verify the accuracy of charts: Ensure that the charts accurately reflect the fluctuations in logistics demand, and make adjustments when necessary to improve the readability of the charts and the accuracy of information expression. Through the above steps, we can clearly show the fluctuations in logistics demand, thereby providing strong support for logistics management and decision-making.

[0078] Step S5, drawing a logistics demand trend curve and a logistics demand fluctuation pattern curve based on the visualization chart.

[0079] Specifically, based on the visualization chart: the premise of this step is that a visualization chart that can intuitively display the fluctuation of logistics demand has been generated. These charts are obtained by comprehensive analysis of multi-dimensional analysis results and logistics extraction features. Draw the logistics demand trend curve: In this link, we will further refine the changing trend of logistics demand over time from a visualization perspective. The trend curve can help us observe the long-term change direction of logistics demand, such as growth, decline or stability. Logistics demand fluctuation pattern curve: In addition to the trend curve, we also need to draw the logistics demand fluctuation pattern curve. This curve focuses on the short-term change pattern of logistics demand, such as seasonal fluctuations, cyclical changes, etc. It helps us understand the changing law of logistics demand within a specific time range.

[0080] Specifically, we can implement this solution in the following steps: Select appropriate data points: Select key data points representing logistics demand from the previous multi-dimensional analysis results. These data points should cover different time periods and different aspects of logistics activities. Determine the time series: Construct a time series based on the selected data points to ensure that the data is arranged in chronological order for subsequent trend and fluctuation analysis. Calculate the trend line: Use appropriate methods (such as moving average method, exponential smoothing method, etc.) to calculate the trend line of logistics demand. The trend line should reflect the overall direction of logistics demand changes over time. Analyze the fluctuation pattern: Identify the fluctuation pattern of logistics demand through further analysis of the time series data. This may involve technical means such as periodic analysis and seasonal decomposition. Draw a trend curve: Use chart tools (such as Excel, Matplotlib or Seaborn in Python, etc.) to draw the trend curve of logistics demand. This curve should clearly show the long-term trend of logistics demand. Draw the fluctuation pattern curve: Use the same chart tool to draw the fluctuation pattern curve of logistics demand. This curve should highlight the short-term changes in logistics demand, such as periodic peaks and troughs. Comprehensive display: Put the trend curve and the fluctuation pattern curve on the same chart, or use different charts to display the two curves separately, so that users can observe the long-term trend and short-term fluctuation pattern of logistics demand at the same time. Optimization and explanation: Finally, optimize the drawn curve to ensure that it is easy to understand, and add necessary annotations to explain the logistics demand characteristics reflected by the curve. Through such steps, we can effectively display the trend and fluctuation pattern of logistics demand, and provide strong support for logistics planning and decision-making.

[0081] Step S6, using a preset prediction algorithm, perform logistics type prediction based on the logistics demand trend curve and the logistics demand fluctuation pattern curve to obtain the type of logistics demand, and formulate a corresponding logistics demand strategy based on the type of logistics demand.

[0082] Specifically, the preset prediction algorithm: here refers to the pre-selected statistical or machine learning method used to predict the changes in future logistics demand. Logistics demand trend curve: refers to the curve drawn in the previous step that shows the long-term trend of logistics demand. Logistics demand fluctuation pattern curve: refers to the curve drawn in the previous step that shows the short-term fluctuation pattern of logistics demand. Logistics type prediction: In this step, we will combine the information of the trend curve and the fluctuation pattern curve to predict the future logistics demand type. The logistics demand type includes but is not limited to the demand for different types of goods, the time distribution of demand, etc. The type of logistics demand: refers to the specific classification of logistics demand determined according to the prediction results. These classifications can guide the formulation of logistics management strategies. Logistics demand strategy: Based on the predicted logistics demand type, formulate corresponding strategies to meet these demands, improve logistics efficiency and reduce costs. The specific steps are as follows: Select a prediction algorithm: Select a suitable prediction model based on the previous analysis and data characteristics. Common prediction algorithms include but are not limited to time series analysis, ARIMA model, exponential smoothing method, neural network model, etc. Use trend curve and fluctuation pattern curve to train the model: The logistics demand trend curve and logistics demand fluctuation pattern curve are used as input data to train the selected prediction algorithm to establish a prediction model. Logistics type prediction: Use the trained model to predict the logistics demand in the future. The prediction results should include the expected changes of different types of logistics demand. Evaluate the prediction accuracy: Verify the prediction results. The accuracy and reliability of the prediction model can be evaluated through historical data backtesting. Determine the type of logistics demand: Based on the prediction results, clarify different types of logistics demand, such as fast-moving consumer goods logistics, bulk cargo transportation, etc. Formulate logistics demand strategies: Formulate specific logistics strategies for each type of logistics demand. For example, for goods with obvious seasonal fluctuations, you can reserve transportation capacity in advance; for goods with stable growth, you can consider increasing warehouse capacity or optimizing distribution routes. Implementation and monitoring: Put the formulated strategies into practice, and continuously monitor the effects of logistics operations, and adjust the strategies according to actual conditions when necessary. Feedback and optimization: Collect data from the actual execution process, feed it back to the prediction model, and continuously optimize the prediction algorithm and logistics demand strategy. Through the above steps, the prediction algorithm and the existing trend curves and fluctuation pattern curves can be effectively used to predict the future logistics demand types, and formulate reasonable logistics demand strategies based on them, thereby improving the overall efficiency and response speed of the logistics system.

[0083] In a specific embodiment, multiple data sources in the logistics supply chain are collected to obtain multi-source collected data, including:

[0084] Collect multiple data sources in the logistics supply chain to obtain collected data;

[0085] Performing standardization processing on the collected data to obtain standardized data;

[0086] Identifying and removing duplicates from the standardized data to obtain identified and removed duplicates data;

[0087] The identified and deduplicated data are classified to obtain multi-source collected data.

[0088] Specifically, multiple data sources in the logistics supply chain are collected to obtain the collected data:

[0089] Multiple data sources in the logistics supply chain: This usually includes various systems and records in the entire chain from suppliers, manufacturers, distributors, retailers to end consumers. Data sources may include ERP (Enterprise Resource Planning), WMS (Warehouse Management System), TMS (Transportation Management System), POS (Point of Sale System), etc. Collect data: Obtain raw data from the various systems mentioned above through interface integration, file transfer, etc. These data include order information, inventory status, transportation records, etc. Standardize the collected data to obtain standardized data: Standardization: Since the data comes from different systems and sources, there may be problems such as inconsistent formats and inconsistent units. Standardization includes operations such as cleaning data, converting data formats, and unifying measurement units to ensure that all data are comparable. Standardized data: After standardization, the data is now consistent in format and unit, and can be further analyzed. Identify and deduplicate the standardized data to obtain identified and deduplicated data: Identification and deduplication: Since the data in the logistics supply chain may have duplicate records, it is necessary to use certain algorithms or rules to identify duplicate data items and delete or merge them to avoid errors in subsequent analysis. Identify and remove duplicate data: After identification and deduplication, the data is unique and accurate. Classify the identified and deduplicated data to obtain multi-source collected data: Classification: Divide the data into different categories according to the attributes or characteristics of the data. For example, it can be classified by the source, type or business function of the data. Multi-source collected data: The data set obtained after classification has been classified and organized according to its attributes for subsequent analysis and use. The specific steps are as follows: Data source collection: Collect raw data from different links of the logistics supply chain, including but not limited to order information, inventory levels, transportation records, etc. Data standardization: Clean and convert the collected data, unify the format and unit, and ensure the consistency and comparability of the data. Data identification and deduplication: Use data deduplication technology to remove duplicate records to ensure that each record in the data set is unique. Data classification: Classify and organize the data according to business needs and the purpose of data analysis, such as by product category, geographic location or business process. Through the above steps, we can obtain structured, high-quality and clearly classified multi-source collected data, which can be used to support decision analysis, predictive modeling and other advanced data analysis tasks in the logistics supply chain.

[0090] In a specific embodiment, a preset principal component analysis algorithm is used to extract logistics features from the multi-source collected data to obtain logistics extraction features, including:

[0091] Performing logistics feature extraction on the multi-source collected data to obtain preliminary extracted features, and performing vector conversion on the preliminary extracted features to obtain a multi-source collected vector;

[0092] Inputting the multi-source acquisition vector into a preset principal component analysis algorithm to perform principal component analysis to obtain a principal component analysis result;

[0093] Constructing a feature matrix based on the principal component analysis result, and calculating a covariance matrix based on the feature matrix; wherein the covariance matrix is ​​used to describe the linear relationship between each logistics extraction feature and the strength of the logistics extraction feature;

[0094] Performing eigenvalue decomposition on the covariance matrix to obtain principal component eigenvectors and eigenvalues ​​corresponding to the principal component eigenvectors;

[0095] Calculating the variance explanation ratio of the principal component analysis result based on the principal component eigenvector and the eigenvalue corresponding to the principal component eigenvector to evaluate the importance of each principal component analysis result in the multi-source collected data and obtain the variance explanation power of each principal component analysis result;

[0096] By using a preset load weight extraction algorithm, based on each of the variance explanatory power, load weight calculation is performed on the preliminary extracted features to obtain a weight coefficient of each preliminary extracted feature;

[0097] Setting a weight coefficient threshold, removing the weight coefficients of each initially extracted feature that is lower than the weight coefficient threshold, and obtaining a standard weight coefficient;

[0098] The standard weight coefficient is matched with the logistics extraction feature to obtain the corresponding logistics extraction feature.

[0099] Specifically, the multi-source collected data is subjected to logistics feature extraction to obtain preliminary extracted features, and the preliminary extracted features are subjected to vector transformation to obtain a multi-source collected vector: Logistics feature extraction: The goal of this step is to extract features related to logistics demand forecasting from multi-source collected data. These features may include order quantity, inventory level, transportation time, etc. Vector transformation: The extracted features are transformed into a numerical vector form to facilitate mathematical operations and analysis. For example, the value of each feature is formed into a vector. The multi-source collected vector is input into a preset principal component analysis algorithm for principal component analysis to obtain the principal component analysis result: Principal component analysis (PCA): This is a commonly used dimensionality reduction technique for extracting new features (i.e., principal components) from original features. These principal components are linear combinations of the original features and are orthogonal to each other. Principal component analysis result: The result obtained by the PCA algorithm includes the principal components and their corresponding variance contribution rates. The principal component analysis result is subjected to the construction of a feature matrix, and the covariance matrix is ​​calculated based on the feature matrix: Construction of feature matrix: The extracted preliminary features are formed into a matrix, each row represents an observation sample, and each column represents a feature. Calculate the covariance matrix: The covariance matrix is ​​used to describe the linear relationship between each logistics extraction feature and the strength of the logistics extraction feature. It is an important part of the PCA algorithm and is used to calculate the principal component. Perform eigenvalue decomposition on the covariance matrix to obtain the principal component eigenvector and the eigenvalue corresponding to the principal component eigenvector: Eigenvalue decomposition: This is a mathematical operation used to decompose the covariance matrix into eigenvalues ​​and eigenvectors. The eigenvalue indicates the variance of each principal component, and the eigenvector indicates the direction of the principal component. Principal component eigenvector and eigenvalue: The eigenvector represents the main direction of the original data, and the eigenvalue indicates the variance along these directions. Based on the principal component eigenvector and the eigenvalue corresponding to the principal component eigenvector, calculate the variance explanation ratio of the principal component analysis result to evaluate the importance of each principal component analysis result in the multi-source acquisition data, and obtain the variance explanation power of each principal component analysis result: Variance explanation ratio: This is an indicator of the ability of each principal component to explain the variance of the original data, usually expressed as a percentage. A higher variance explanation ratio means that the principal component can better explain the variation of the data. Variance Explanation Power: The variance explanatory power of each principal component analysis result reflects the importance of the principal component in the original data. Through the preset load weight extraction algorithm, the load weight of the preliminary extracted features is calculated based on each of the variance explanatory powers to obtain the weight coefficient of each preliminary extracted feature: Load weight extraction algorithm: This is an algorithm for calculating the weight of the preliminary extracted features based on the variance explanatory power of each principal component. Weight coefficient: The weight coefficient represents the relative importance of the preliminary extracted features in the principal component analysis results.Set a weight coefficient threshold, remove the weight coefficients of each preliminary extracted feature that is lower than the weight coefficient threshold, and obtain a standard weight coefficient: Weight coefficient threshold: This is a predetermined value used to filter out unimportant features. Standard weight coefficient: The retained weight coefficients, the important features corresponding to these coefficients will be used for subsequent analysis. Perform logistics extraction feature matching on the standard weight coefficient to obtain the corresponding logistics extraction features: Logistics extraction feature matching: According to the standard weight coefficient, select those features with larger weights as the final logistics extraction features. Logistics extraction features: These features are important features extracted from the original data and can be used for subsequent logistics demand forecasting analysis. Through the above steps, we can extract key logistics features from multi-source collected data, and perform dimensionality reduction processing on these features through the principal component analysis algorithm, and finally obtain a set of logistics extraction features that can represent the main information of the original data. These features can be used for subsequent logistics demand forecasting analysis.

[0100] In a specific embodiment, a multi-dimensional analysis is performed on the logistics extraction features to obtain a multi-dimensional analysis result, including:

[0101] Performing a multi-dimensional analysis on the logistics extraction features through a preset hidden Markov model to obtain a preliminary multi-dimensional analysis result;

[0102] Based on the multi-dimensional analysis results, the hidden patterns of logistics demand are identified to reveal the driving factors of changes in logistics demand;

[0103] Using a preset Bayesian network, conditional dependency inference is performed on the logistics extraction features based on the driving factors to obtain an inference result;

[0104] Using a preset interactive visualization technology, the correlation lines between the inference result and the preliminary multi-dimensional analysis result are displayed to obtain a conditional dependency relationship diagram;

[0105] The conditional dependency diagram is interpreted to obtain multi-dimensional analysis results.

[0106] Specifically, the logistics extraction features are subjected to multi-dimensional analysis by a preset hidden Markov model to obtain preliminary multi-dimensional analysis results: Hidden Markov Model (HMM): This is a statistical model that is often used to analyze data with time series properties. Here, HMM is used to capture the change pattern of logistics extraction features over time. Preliminary multi-dimensional analysis results: The logistics extraction features are analyzed by HMM to obtain preliminary analysis results, which may include the relationship between features, the change trend of features over time, etc. Based on the multi-dimensional analysis results, the hidden patterns of logistics demand are identified to reveal the driving factors of changes in logistics demand: Identify hidden patterns: Based on the preliminary multi-dimensional analysis results, further search for potential patterns behind the changes in logistics demand, which may be caused by some factors that are not directly observed. Driving factors: By analyzing the hidden patterns found, the key factors affecting the changes in logistics demand are identified. Using the preset Bayesian network, the logistics extraction features are subjected to conditional dependency inference based on the driving factors to obtain the inference results: Bayesian network: This is a probabilistic graph model used to represent the conditional dependency between variables. Here, the Bayesian network is used to infer the conditional dependency between logistics extraction features. Conditional dependency inference: Based on the identified driving factors, the conditional dependency between different features is inferred using a Bayesian network. Inference results: The conditional dependency obtained by the Bayesian network. The preset interactive visualization technology is used to display the association lines between the inference results and the preliminary multidimensional analysis results to obtain a conditional dependency graph: Interactive visualization technology: This is a visualization method that allows users to explore and manipulate data through a graphical interface to better understand the relationship between data. Conditional dependency graph: The conditional dependency between logistics extraction features and the connection between these relationships and the preliminary multidimensional analysis results are displayed through interactive visualization technology. The conditional dependency graph is interpreted to obtain multidimensional analysis results: Interpretation: Through the analysis of the conditional dependency graph, the relationship between logistics extraction features and how these relationships affect logistics demand are understood. Multidimensional analysis results: The comprehensive analysis results obtained through the above steps include the change pattern of logistics demand, driving factors, and conditional dependencies between features. Through the above steps, we can reveal the complex patterns and driving factors behind the changes in logistics demand from the logistics extraction features, and intuitively display these relationships through the conditional dependency graph. These multidimensional analysis results provide valuable information for subsequent logistics demand forecasting.

[0107] In a specific embodiment, logistics demand is identified based on the multi-dimensional analysis results and logistics extraction features to obtain the fluctuation of logistics demand, and a visualization chart is generated based on the fluctuation, including:

[0108] Based on the multi-dimensional analysis results, the logistics demand is identified to obtain logistics demand information; wherein the logistics demand information includes demand quantity, demand frequency, demand cycle and demand urgency;

[0109] quantifying the logistics demand information to obtain the fluctuation range of the logistics demand information;

[0110] Using the preset data genomics technology to perform pattern recognition on the logistics extraction features, the fundamental characteristics of logistics demand are obtained;

[0111] Based on the fundamental characteristics and the fluctuation range, the logistics demand is deeply analyzed to obtain the fluctuation of the logistics demand;

[0112] Determine whether there are abnormal fluctuations in the fluctuation situation. If so, visualize the fluctuation situation of the logistics demand based on the abnormal fluctuations to obtain a dynamic fluctuation curve graph, and use the dynamic fluctuation curve graph as a visualization chart; wherein, the abnormal fluctuations include a surge in demand, a sudden drop in demand and abnormal periodic fluctuations.

[0113] Specifically, in the process of logistics demand identification and analysis, it is necessary to first conduct a comprehensive analysis of the multi-dimensional analysis results to identify the fluctuations in logistics demand. The core of logistics demand identification lies in extracting and analyzing relevant features from huge data. Through these features, we can obtain key information such as demand volume, demand frequency, demand cycle, and demand urgency. The accurate acquisition of this information is crucial to the efficiency of logistics operations. Therefore, on the basis of multi-dimensional analysis, it is necessary to first deeply identify these demand information. In the process of identifying logistics demand information, the demand volume must be considered first, which determines the number of goods or the frequency of services that need to be transported. Demand frequency is an analysis of the manifestation of demand in a certain time period, determining the frequency of demand and whether it is regular. The identification of demand cycles is more complicated, and it is necessary to take into account the seasonal changes in market demand, holiday fluctuations, and the impact of market trends. Finally, these factors are combined to identify the performance of logistics demand in different cycles. The urgency of demand is mainly related to the restriction and urgency of order delivery time, which directly affects the operation speed of each link in the logistics chain. After obtaining and identifying this information, the next step is to quantify this logistics demand information. The quantification process is mainly to convert the demand, frequency, cycle and urgency into measurable values, which is convenient for calculation and processing in subsequent analysis. After quantification, we will get the fluctuation range of demand information, which is an important indicator to describe the change of logistics demand. It can reveal the upward or downward trend of logistics demand within a certain time range. Then, the preset data genomics technology is used to perform pattern recognition on the logistics extraction features. The key to this step is to identify the fundamental characteristics that affect the change of logistics demand through big data analysis technology. Data genomics technology can be understood as an efficient algorithm or technical means. It can extract representative patterns or features from massive data. These features are not only a description of current logistics demand, but also can help predict future demand changes. Through this technical means, it is possible to identify the potential laws hidden behind the data and reveal the fundamental characteristics of logistics demand. Based on these fundamental characteristics, combined with the fluctuation range obtained above, a more in-depth analysis of logistics demand is carried out. In this step, we need to comprehensively consider the fundamental characteristics and fluctuation range of logistics demand to analyze the overall fluctuation of logistics demand. This process is not only the processing of existing data, but also involves the prediction of future demand changes. Through in-depth analysis, it can be concluded whether there are abnormal fluctuations in logistics demand, such as a surge in demand, a sudden drop in demand, or abnormal cyclical fluctuations. Once these abnormal fluctuations are identified, the system will further visualize the fluctuations in logistics demand based on these abnormalities. In this process, dynamic fluctuation curves are used to intuitively display the trend of demand changes. These curves can not only reflect changes in demand in real time, but also help logistics managers make quick response decisions.Through these dynamic charts, managers can clearly see the growth or decrease trend of demand, so as to adjust resource allocation in time and ensure the efficient operation of the logistics system. In short, this series of steps starts with multi-dimensional analysis, through the identification of demand information, quantitative processing, and pattern recognition of fundamental characteristics, and finally achieves in-depth analysis of logistics demand and presents it through visualization tools. These steps are interrelated to form a complete analysis and decision support system to ensure that logistics demand can be accurately identified and responded to. In logistics demand analysis, "fundamental characteristics" refer to key attributes or patterns extracted from a large amount of logistics data through data genomics technology. These characteristics can reflect the core factors of changes in logistics demand. The identification of fundamental characteristics is the core step of data analysis. It helps us understand the most basic driving factors affecting logistics demand, thereby providing strong support in the decision-making process.

[0114] Fundamental characteristics usually include but are not limited to the following aspects: Demand pattern characteristics: This refers to the changing rules of demand in the time dimension, such as seasonal fluctuations in demand, differences between weekends and weekdays, peak periods during holidays, etc. These characteristics can help identify cyclical changes in demand and provide a basis for resource allocation. Geographic distribution characteristics: Logistics demand in different regions may be significantly different. Fundamental characteristics include the distribution of demand in different geographical regions, such as differences in demand between urban and rural areas, and between eastern and western regions. These characteristics can help optimize the layout of the logistics network and improve logistics efficiency. Customer behavior characteristics: Customers' purchasing habits and behavior patterns are also important drivers of logistics demand. For example, behavioral characteristics such as customer order time, order volume, and return rate can all be part of the fundamental characteristics. These characteristics can help identify the demand characteristics of different customer groups and customize logistics services. Market trend characteristics: The overall trend of the market, such as the increase or decrease in demand for a certain type of commodity, the rise of emerging markets, etc., will have an impact on logistics demand. By analyzing these market trend characteristics, future changes in logistics demand can be predicted and prepared in advance. Supply chain fluctuation characteristics: Changes in the upstream supply chain, such as suppliers' supply capacity and fluctuations in raw material prices, will also affect logistics demand. These characteristics can help identify external factors affecting logistics demand and provide a reference for supply chain management. By comprehensively analyzing these fundamental characteristics, we can more accurately understand the reasons for changes in logistics demand and fully consider these core factors in decision-making. Identifying and utilizing fundamental characteristics can greatly improve the response speed and response capabilities of the logistics system, thereby better meeting market demand.

[0115] In a specific embodiment, drawing a logistics demand trend curve and a logistics demand fluctuation pattern curve based on the visualization chart includes:

[0116] Deconstructing the data in the visualization chart to obtain deconstructed data;

[0117] Identifying long-term and / or short-term trends from the deconstructed data; using the long-term and / or short-term trends as logistics demand trends;

[0118] Draw a logistics demand trend curve based on the logistics demand trend;

[0119] The visualization chart is interpreted by Fourier transform to obtain the frequency components of the demand changes; wherein the frequency components include fundamental frequency and harmonics;

[0120] Analyzing the amplitudes of the different frequency components to obtain amplitude analysis results;

[0121] Constructing a frequency spectrum based on the amplitude analysis result, and obtaining a fluctuation pattern of each frequency component to logistics demand based on the frequency spectrum;

[0122] A logistics demand fluctuation pattern curve is drawn based on the logistics demand fluctuation pattern.

[0123] Specifically, the data in the visualization chart is deconstructed to obtain deconstructed data: Deconstructing data: splitting the data in the visualization chart into smaller, analyzable parts to facilitate subsequent analysis.

[0124] Identify long-term and / or short-term trends from the deconstructed data; use the long-term and / or short-term trends as logistics demand trends: Long-term trend: refers to the overall change direction of logistics demand over a long period of time. Short-term trend: refers to the changing characteristics of logistics demand within a short period of time. Logistics demand trend: integrate long-term and / or short-term trends to form an overall understanding of the changing trend of logistics demand. Logistics demand trend curve: display the changing trend of logistics demand over time in the form of a chart, and intuitively reflect the long-term and / or short-term trend of demand. Use Fourier transform to interpret the visualization chart to obtain the frequency components of demand changes: Fourier transform: a mathematical method for converting time series data into data in the frequency domain, thereby revealing the periodicity and frequency characteristics in the data. Frequency components: different frequency components in the change of logistics demand obtained by Fourier transform, including fundamental frequency and harmonics. Analyze the amplitudes of different frequency components to obtain amplitude analysis results: Amplitude analysis results: analyze the intensity or amplitude of each frequency component to understand their impact on the overall demand change. A frequency spectrum is constructed based on the amplitude analysis results, and the logistics demand fluctuation pattern of each frequency component is obtained based on the frequency spectrum: Frequency spectrum: A chart used to display the amplitude distribution of different frequency components to help identify the main frequency components. Logistics demand fluctuation pattern: By analyzing the frequency spectrum, the main patterns and periodic characteristics in the changes in logistics demand are identified. A logistics demand fluctuation pattern curve is drawn based on the logistics demand fluctuation pattern: Logistics demand fluctuation pattern curve: A curve drawn according to the logistics demand fluctuation pattern to intuitively display the periodic fluctuations and patterns in demand changes. Through the above steps, we can have a deeper understanding of the time series characteristics of logistics demand, especially the periodic and trend changes therein. These analysis results can help companies better predict future logistics demand, and optimize logistics resource allocation and improve operational efficiency accordingly.

[0125] In a specific embodiment, the type of logistics demand trend and logistics demand fluctuation pattern is predicted by a preset prediction algorithm to obtain the type of logistics demand, and a corresponding logistics demand strategy is formulated based on the type of logistics demand, including:

[0126] The logistics demand trend is predicted by a preset prediction algorithm to obtain the dynamic demand of logistics demand; wherein the dynamic demand includes seasonal fluctuations, market penetration rate and supply chain dynamics;

[0127] Predicting the logistics demand fluctuation pattern to obtain a demand change pattern; wherein the demand change pattern includes a stable pattern, a growth pattern, a fluctuation pattern and a sudden pattern;

[0128] Based on the dynamic demand of the logistics demand and the demand change pattern, the type of the logistics demand is predicted to obtain the type of the logistics demand;

[0129] Formulate a preliminary logistics strategy based on the type of logistics demand; wherein the preliminary logistics strategy includes inventory management, transportation planning and resource allocation;

[0130] Conducting risk assessment on the implementation process of the preliminary logistics strategy to obtain risk assessment results;

[0131] The preliminary logistics strategy is optimized based on the risk assessment result to obtain a corresponding logistics demand strategy.

[0132] Specifically, the logistics demand trend is predicted by a preset prediction algorithm to obtain the dynamic demand of logistics demand: Dynamic demand: refers to the volatility characteristics of logistics demand over time, including seasonal fluctuations, market penetration and supply chain dynamics. The logistics demand fluctuation pattern is predicted to obtain the demand change pattern: Demand change pattern: The way of predicting future demand changes by analyzing historical data can be divided into several types such as stable mode, growth mode, fluctuation mode and sudden mode. Based on the dynamic demand of the logistics demand and the demand change pattern, the type of logistics demand is predicted to obtain the type of logistics demand: Type of logistics demand: According to the combination of dynamic demand and demand change pattern, the specific type of logistics demand is predicted and defined, which is helpful to formulate targeted strategies. A preliminary logistics strategy is formulated for the type of logistics demand: Preliminary logistics strategy: The basic strategy framework formulated according to the type of logistics demand, including inventory management, transportation planning and resource allocation. The implementation process risk assessment of the preliminary logistics strategy is carried out to obtain the risk assessment result: Implementation process risk assessment: The risks that may be encountered in the actual operation of the preliminary logistics strategy are assessed. Risk assessment result: The risk level obtained after the assessment and its impact on the logistics strategy. Based on the risk assessment results, the preliminary logistics strategy is optimized to obtain the corresponding logistics demand strategy: Strategy optimization: Adjust the preliminary logistics strategy according to the risk assessment results to reduce potential risks and ensure the effectiveness and feasibility of the strategy. Corresponding logistics demand strategy: The final logistics strategy after optimization can better adapt to changes in logistics demand and effectively deal with various risks. Through the above steps, enterprises can formulate more reasonable and effective logistics strategies based on the prediction of future logistics demand trends and fluctuation patterns, thereby improving the efficiency and response speed of logistics operations, reducing costs and improving customer satisfaction.

[0133] The above describes the logistics demand prediction method based on big data analysis in the embodiment of the present invention. The following describes the logistics demand prediction device based on big data analysis in the embodiment of the present invention. Figure 2 In one embodiment of the present invention, a logistics demand prediction device based on big data analysis includes:

[0134] The collection module 21 is used to collect multiple data sources in the logistics supply chain to obtain multi-source collected data; wherein the multi-source collected data includes order information, inventory levels and transportation records;

[0135] An extraction module 22 is used to extract logistics features from the multi-source collected data using a preset principal component analysis algorithm to obtain logistics extraction features; wherein the logistics extraction features include demand fluctuations and time series trends;

[0136] The analysis module 23 is used to perform multi-dimensional analysis on the logistics extraction features to obtain multi-dimensional analysis results; wherein the multi-dimensional analysis results include demand fluctuation frequency, seasonal trend analysis, regional demand difference analysis, product category demand changes, customer demand behavior patterns and transportation mode efficiency comparison;

[0137] An identification module 24 is used to identify logistics demand based on the multi-dimensional analysis results and logistics extraction features, obtain the fluctuation of logistics demand, and generate a visual chart based on the fluctuation;

[0138] A drawing module 25, used for drawing a logistics demand trend curve and a logistics demand fluctuation pattern curve based on the visualization chart;

[0139] The prediction module 26 is used to predict the logistics type based on the logistics demand trend curve and the logistics demand fluctuation pattern curve through a preset prediction algorithm, obtain the type of logistics demand, and formulate a corresponding logistics demand strategy based on the type of logistics demand.

[0140] In this embodiment, for the specific implementation of each unit in the above system embodiment, please refer to the above method embodiment, which will not be repeated here.

[0141] Reference Figure 3 The present invention also provides a computer device in an embodiment, wherein the internal structure of the computer device can be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a display screen, an input device, a network interface and a database connected through a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the above method is implemented.

[0142] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

[0143] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the above method is implemented. It can be understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0144] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided by the present invention and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM.

[0145] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0146] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A logistics demand forecasting method based on big data analysis, characterized in that: The following steps are involved: Collect multiple data sources in the logistics supply chain to obtain multi-source collected data; wherein the multi-source collected data includes order information, inventory levels and transportation records; Using a preset principal component analysis algorithm to extract logistics features from the multi-source collected data, to obtain logistics extraction features; wherein the logistics extraction features include demand fluctuations and time series trends; Performing a multi-dimensional analysis on the logistics extraction features to obtain a multi-dimensional analysis result; wherein the multi-dimensional analysis result includes demand fluctuation frequency, seasonal trend analysis, regional demand difference analysis, product category demand changes, customer demand behavior patterns and transportation mode efficiency comparison; Based on the multi-dimensional analysis results and the logistics extraction features, the logistics demand is identified to obtain the fluctuation of the logistics demand, and a visual chart is generated based on the fluctuation; Draw a logistics demand trend curve and a logistics demand fluctuation pattern curve based on the visualization chart; By using a preset prediction algorithm, a logistics type prediction is performed based on the logistics demand trend curve and the logistics demand fluctuation pattern curve to obtain the type of logistics demand, and a corresponding logistics demand strategy is formulated based on the type of logistics demand; Drawing a logistics demand trend curve and a logistics demand fluctuation pattern curve based on the visualization chart includes: Deconstructing the data in the visualization chart to obtain deconstructed data; Identifying long-term and / or short-term trends from the deconstructed data; using the long-term and / or short-term trends as logistics demand trends; Draw a logistics demand trend curve based on the logistics demand trend; The visualization chart is interpreted by Fourier transform to obtain the frequency components of the demand changes; wherein the frequency components include fundamental frequency and harmonics; Analyzing the amplitudes of the different frequency components to obtain amplitude analysis results; Constructing a frequency spectrum based on the amplitude analysis result, and obtaining a fluctuation pattern of each frequency component to logistics demand based on the frequency spectrum; A logistics demand fluctuation pattern curve is drawn based on the logistics demand fluctuation pattern.

2. The logistics demand forecasting method based on big data analysis according to claim 1 is characterized in that: Collect multiple data sources in the logistics supply chain to obtain multi-source data, including: Collect multiple data sources in the logistics supply chain to obtain collected data; Performing standardization processing on the collected data to obtain standardized data; Identifying and removing duplicates from the standardized data to obtain identified and removed duplicates data; The identified and deduplicated data are classified to obtain multi-source collected data.

3. The logistics demand forecasting method based on big data analysis according to claim 1 is characterized in that: The logistics features of the multi-source collected data are extracted using a preset principal component analysis algorithm to obtain logistics extraction features, including: Performing logistics feature extraction on the multi-source collected data to obtain preliminary extracted features, and performing vector conversion on the preliminary extracted features to obtain a multi-source collected vector; Inputting the multi-source acquisition vector into a preset principal component analysis algorithm to perform principal component analysis to obtain a principal component analysis result; Constructing a feature matrix based on the principal component analysis result, and calculating a covariance matrix based on the feature matrix; wherein the covariance matrix is ​​used to describe the linear relationship between each logistics extraction feature and the strength of the logistics extraction feature; Performing eigenvalue decomposition on the covariance matrix to obtain principal component eigenvectors and eigenvalues ​​corresponding to the principal component eigenvectors; Calculating the variance explanation ratio of the principal component analysis result based on the principal component eigenvector and the eigenvalue corresponding to the principal component eigenvector to evaluate the importance of each principal component analysis result in the multi-source collected data and obtain the variance explanation power of each principal component analysis result; By using a preset load weight extraction algorithm, based on each of the variance explanatory power, load weight calculation is performed on the preliminary extracted features to obtain weight coefficients of each preliminary extracted feature; Setting a weight coefficient threshold, removing the weight coefficients of each initially extracted feature that is lower than the weight coefficient threshold, and obtaining a standard weight coefficient; The standard weight coefficient is matched with the logistics extraction feature to obtain the corresponding logistics extraction feature.

4. The logistics demand forecasting method based on big data analysis according to claim 1 is characterized in that: The logistics extraction features are subjected to multi-dimensional analysis to obtain multi-dimensional analysis results, including: Performing a multi-dimensional analysis on the logistics extraction features through a preset hidden Markov model to obtain a preliminary multi-dimensional analysis result; Based on the multi-dimensional analysis results, the hidden patterns of logistics demand are identified to reveal the driving factors of changes in logistics demand; Using a preset Bayesian network, conditional dependency inference is performed on the logistics extraction features based on the driving factors to obtain an inference result; Using a preset interactive visualization technology, the correlation lines between the inference result and the preliminary multi-dimensional analysis result are displayed to obtain a conditional dependency relationship diagram; The conditional dependency diagram is interpreted to obtain multi-dimensional analysis results.

5. The logistics demand forecasting method based on big data analysis according to claim 1 is characterized in that: Based on the multi-dimensional analysis results and the logistics extraction features, the logistics demand is identified to obtain the fluctuation of the logistics demand, and a visualization chart is generated based on the fluctuation, including: Based on the multi-dimensional analysis results, the logistics demand is identified to obtain logistics demand information; wherein the logistics demand information includes demand quantity, demand frequency, demand cycle and demand urgency; quantifying the logistics demand information to obtain the fluctuation range of the logistics demand information; Using the preset data genomics technology to perform pattern recognition on the logistics extraction features, the fundamental characteristics of logistics demand are obtained; Based on the fundamental characteristics and the fluctuation range, the logistics demand is deeply analyzed to obtain the fluctuation of the logistics demand; Determine whether there are abnormal fluctuations in the fluctuation situation. If so, visualize the fluctuation situation of the logistics demand based on the abnormal fluctuations to obtain a dynamic fluctuation curve graph, and use the dynamic fluctuation curve graph as a visualization chart; wherein, the abnormal fluctuations include a surge in demand, a sudden drop in demand and abnormal cyclical fluctuations.

6. The logistics demand forecasting method based on big data analysis according to claim 1 is characterized in that: The logistics demand trend and the logistics demand fluctuation pattern are predicted by a preset prediction algorithm to obtain the type of logistics demand, and a corresponding logistics demand strategy is formulated based on the type of logistics demand, including: The logistics demand trend is predicted by a preset prediction algorithm to obtain the dynamic demand of logistics demand; wherein the dynamic demand includes seasonal fluctuations, market penetration rate and supply chain dynamics; Predicting the logistics demand fluctuation pattern to obtain a demand change pattern; wherein the demand change pattern includes a stable pattern, a growth pattern, a fluctuation pattern and a sudden pattern; Based on the dynamic demand of the logistics demand and the demand change pattern, the type of the logistics demand is predicted to obtain the type of the logistics demand; Formulate a preliminary logistics strategy based on the type of logistics demand; wherein the preliminary logistics strategy includes inventory management, transportation planning and resource allocation; Conducting risk assessment on the implementation process of the preliminary logistics strategy to obtain risk assessment results; The preliminary logistics strategy is optimized based on the risk assessment result to obtain a corresponding logistics demand strategy.

7. A logistics demand forecasting device based on big data analysis, characterized in that: Applied to five-axis laser equipment, including: A collection module is used to collect multiple data sources in the logistics supply chain to obtain multi-source collected data; wherein the multi-source collected data includes order information, inventory levels and transportation records; An extraction module, used to extract logistics features from the multi-source collected data using a preset principal component analysis algorithm to obtain logistics extraction features; wherein the logistics extraction features include demand fluctuations and time series trends; An analysis module, used to perform multi-dimensional analysis on the logistics extraction features to obtain multi-dimensional analysis results; wherein the multi-dimensional analysis results include demand fluctuation frequency, seasonal trend analysis, regional demand difference analysis, product category demand changes, customer demand behavior patterns and transportation mode efficiency comparison; An identification module, used to identify logistics demand based on the multi-dimensional analysis results and logistics extraction features, obtain the fluctuation of logistics demand, and generate a visual chart based on the fluctuation; A drawing module, used for drawing a logistics demand trend curve and a logistics demand fluctuation pattern curve based on the visualization chart; A prediction module, used to predict the logistics type based on the logistics demand trend curve and the logistics demand fluctuation pattern curve through a preset prediction algorithm, obtain the type of logistics demand, and formulate a corresponding logistics demand strategy based on the type of logistics demand; Drawing a logistics demand trend curve and a logistics demand fluctuation pattern curve based on the visualization chart includes: Deconstructing the data in the visualization chart to obtain deconstructed data; Identifying long-term and / or short-term trends from the deconstructed data; using the long-term and / or short-term trends as logistics demand trends; Draw a logistics demand trend curve based on the logistics demand trend; The visualization chart is interpreted by Fourier transform to obtain the frequency components of the demand changes; wherein the frequency components include fundamental frequency and harmonics; Analyzing the amplitudes of the different frequency components to obtain amplitude analysis results; Constructing a frequency spectrum based on the amplitude analysis result, and obtaining a fluctuation pattern of each frequency component to logistics demand based on the frequency spectrum; A logistics demand fluctuation pattern curve is drawn based on the logistics demand fluctuation pattern.

8. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images