Construction method of special agricultural product price prediction model

Abstract

The invention discloses a construction method of a characteristic agricultural product price prediction model, and relates to the technical field of data prediction, and the method comprises the steps: obtaining a characteristic agricultural product data source, marking the characteristic agricultural product data source as each target data source, collecting the metadata of each target data source, and obtaining agricultural product price prediction data, the metadata comprises data source validity data and data source security data; processing according to the agricultural product price prediction data to obtain an agricultural product data source quality index, and correcting a preset initial data source quality evaluation index threshold according to the agricultural product data source quality index to obtain a reference data source quality evaluation index threshold; analyzing based on the agricultural product price prediction data, the data source security data and the data source validity data, establishing an agricultural product data source evaluation model, and processing to obtain each target data source quality evaluation index; and comparing each target data source quality evaluation index with a reference data source quality evaluation index threshold, screening to obtain each qualified data source, and feeding back.

Description

Technical Field

[0001] The present invention relates to the technical field of data prediction, and in particular to a method for constructing a price prediction model for characteristic agricultural products. Background Art

[0002] As the process of global economic integration continues to accelerate, the integration of specialty agricultural product markets with the international market is becoming increasingly deeper. External factors such as the ups and downs of international agricultural product prices, frequent adjustments in trade policies, and fluctuations in exchange rates have exerted a profound and complex impact on domestic specialty agricultural product prices. At the same time, the rapid development of science and technology has led to the rapid development of agricultural production technology, and the output and quality of specialty agricultural products have been significantly improved. However, this has also led to a profound change in the market supply structure.

[0003] The existing method for constructing a price prediction model for specialty agricultural products is mainly achieved through multiple steps such as data collection, integration and cleaning, feature engineering screening and processing, model selection and construction, model training and optimization, and model evaluation and application.

[0004] For example, the invention patent with publication number CN118505268A discloses a method, system, device and storage medium for predicting the import price of agricultural products, establishes an indicator system of factors affecting the import price of agricultural products, integrates historical and real-time multi-source data, and constructs an influencing factor database; unifies the influencing factors into a monthly update frequency; screens out influencing factors that can play a predictive role; uses a preset algorithm to build a prediction model, and uses historical data on import prices and their influencing factors to train the prediction model; and uses the trained prediction model to predict the import price of agricultural products.

[0005] For example, the invention patent with announcement number CN118261637B announces an agricultural product supply forecasting system and method based on an online trading platform. It collects data on historical transaction information, meteorological information and market trend information, cleans and processes the data to ensure data quality; analyzes the data and extracts features, including sales volume, price and seasonality; builds an agricultural product supply forecasting model, uses historical data to train the model, and achieves a more comprehensive agricultural product supply forecast; predicts future supply conditions, generates forecast results, and displays the forecast results. By evaluating the forecast effect of the agricultural product supply forecasting model, it is effectively determined whether the agricultural product supply forecasting model needs to be intelligently optimized; collects actual sales data, compares it with the forecast results, continuously optimizes the model based on feedback information, and continuously improves the forecast accuracy and effect based on actual conditions.

[0006] However, in the process of implementing the technical solution of the invention in the embodiments of the present application, it was found that the above-mentioned technology has at least the following technical problems: in actual applications, due to the wide range of sources of specialty agricultural product data and uneven data quality, there may be problems such as difficulty in data integration, missing or incorrect key information, which in turn affects the reliability and accuracy of the specialty agricultural product price prediction model, causing a large deviation in the final price prediction result. Summary of the invention

[0007] In view of the deficiencies in the prior art, the present invention provides a method for constructing a price prediction model for characteristic agricultural products, which can effectively solve the problems involved in the above-mentioned background technology.

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions: including: obtaining a data source of characteristic agricultural products, marking it as each target data source, collecting metadata of each target data source, and obtaining agricultural product price prediction data, wherein the metadata includes data source validity data and data source security data.

[0009] The agricultural product data source quality index is obtained based on the agricultural product price forecast data processing, and the preset initial data source quality assessment index threshold is corrected based on the agricultural product data source quality index to obtain the reference data source quality assessment index threshold.

[0010] Based on the analysis of agricultural product price forecast data, data source security data and data source validity data, an agricultural product data source evaluation model is established, and the quality evaluation index of each target data source is obtained through processing.

[0011] The quality assessment index of each target data source is compared with the quality assessment index threshold of the reference data source, and qualified data sources are screened and feedback is provided.

[0012] As a further method, the agricultural product price forecast data is obtained, and the specific analysis process is: obtaining the model's expected prediction time range and the number of price changes of each highland barley variety; processing the number of price changes of each highland barley variety to obtain a data update frequency index; and jointly marking the data update frequency index and the model's expected prediction time range as agricultural product price forecast data.

[0013] As a further method, the data source validity data includes the duration of adjacent update intervals; the data source security data includes data transmission volume, data transmission speed, data source access frequency and number of abnormal interruptions.

[0014] As a further method, the agricultural product data source quality index is obtained based on the agricultural product price forecast data processing. The specific analysis process is: obtaining the critical data update frequency index and the critical model expected forecast time range from the agricultural product database; based on the agricultural product price forecast data, a comprehensive analysis is performed to obtain the agricultural product data source quality index, and the agricultural product data source quality index is used to quantify the demand data quality level for price model construction.

[0015] As a further method, the preset initial data source quality assessment index threshold is corrected according to the agricultural product data source quality index to obtain a reference data source quality assessment index threshold. The specific analysis process is: comparing the agricultural product data source quality index with the data source quality assessment index threshold correction parameters corresponding to each agricultural product data source quality index interval preset in the agricultural product database to obtain the data source quality assessment index threshold correction parameter; summing the preset initial data source quality assessment index threshold and the data source quality assessment index threshold correction parameter to obtain the reference data source quality assessment index threshold.

[0016] As a further method, the processing obtains the quality evaluation index of each target data source, and the specific analysis process is: processing the data source validity data and the data update frequency index to obtain the validity index of each target data source; processing the data source security data to obtain the security index of each target data source; based on the validity index of each target data source and the security index of each target data source, comprehensive analysis is performed to obtain the quality evaluation index of each target data source.

[0017] As a further method, the screening obtains each qualified data source and provides feedback. The specific analysis process is: comparing the quality assessment index of each target data source with the reference data source quality assessment index threshold; if the quality assessment index of a target data source is greater than or equal to the reference data source quality assessment index threshold, then the data source is marked as a qualified data source; if the quality assessment index of a target data source is less than the reference data source quality assessment index threshold, then the data source is marked as an unqualified data source; statistics are obtained for each qualified data source and feedback is provided.

[0018] As a further method, the statistics are obtained for each qualified data source and feedback is provided. The specific analysis process is: statistics are performed on the qualified data sources to obtain each qualified data source, original data of each qualified data source is collected, a qualified data source set is constructed, and the qualified data source set is stored in the agricultural product database for constructing an agricultural product price prediction model.

[0019] Compared with the prior art, the embodiments of the present invention have at least the following beneficial effects:

[0020] (1) The present invention provides a method for constructing a price prediction model for specialty agricultural products, which can fully integrate multi-dimensional data source information, cover data related to specialty agricultural products from different channels, and accurately screen out high-quality data from a large number of data sources for model construction. This method effectively improves the quality of input data for the price prediction model, thereby improving the accuracy of price prediction for specialty agricultural products, helping farmers and dealers to reasonably plan production and sales strategies in advance, and reducing market risks caused by price fluctuations.

[0021] (2) The present invention establishes a detailed evaluation system by comprehensively considering the data source validity data and the data source security data, deeply analyzing the key factors such as data accuracy, timeliness, integrity, and security of transmission and storage, and generating scientific and reasonable quality evaluation indicators for each data source, accurately eliminating data sources with poor quality or potential safety hazards, reducing the deviation of prediction results caused by data problems, and ensuring that the prediction results are more valuable for reference.

[0022] (3) The present invention dynamically corrects the threshold of the initial data source quality assessment index through the agricultural product data source quality index, and screens qualified data sources through threshold comparison. It can flexibly adjust the screening criteria according to the actual conditions such as the price fluctuation characteristics of different specialty agricultural products and the rhythm of market supply and demand changes, thereby enhancing the adaptability of the model to the prediction of price trends of various agricultural products.

[0023] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a schematic diagram of the method flow of the present invention.

[0025] Figure 2 It is a schematic diagram of the functional relationship between the target data source effectiveness index and the target data source quality evaluation index of the present invention. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0028] Reference Figure 1 As shown, the first aspect of the present invention provides a method for constructing a characteristic agricultural product price prediction model, including: obtaining a characteristic agricultural product data source, marking it as each target data source, collecting metadata of each target data source, and obtaining agricultural product price prediction data, wherein the metadata includes data source validity data and data source security data.

[0029] Specifically, agricultural product price forecast data is obtained, and the specific analysis process is: obtain the model's expected prediction time range and the number of price changes of each highland barley variety; process the number of price changes of each highland barley variety to obtain the data update frequency index; jointly mark the data update frequency index and the model's expected prediction time range as agricultural product price forecast data.

[0030] In this embodiment, the data source of featured agricultural products comes from multiple channels, such as statistical databases of local agricultural departments, professional agricultural product trading platforms, and data materials released by agricultural product industry research institutions. The featured agricultural product data sources obtained from different channels are marked separately to facilitate the subsequent targeted management and analysis of the featured agricultural product data sources, and the marked featured agricultural product data sources are identified as target data sources. Metadata is data used to describe data, and metadata includes data source validity data and data source security data.

[0031] It should be understood that in this embodiment, the model is obtained to predict the time range for the prediction of the price of the featured agricultural products. For example, if you want to predict the price trend of a certain featured fruit in the next three months, the predicted time range is 90 days. Collect the number of price changes of each highland barley variety within the predicted time range period, process the price of the featured agricultural products with the number of price changes of each highland barley variety within the period, and obtain the data update frequency index. The data update frequency index is an indicator for measuring the changes in the price of featured agricultural products. By processing the price of the featured agricultural products with the number of price changes of each highland barley variety within the period, an index that can reflect the price fluctuations is obtained, which can intuitively show the fluctuation range of the featured agricultural products in the past period of time. The data update frequency index is combined with the set model predicted prediction time range to form the relevant data for predicting the price of featured agricultural products. The data update frequency index can help analyze the historical laws of agricultural product price changes, and the predicted time range clarifies the specific target interval of the prediction. The two can be used to train the prediction model to predict the trend of the types of featured agricultural products in the future.

[0032] Specifically, the data source validity data includes the duration of the adjacent update interval; the data source security data includes the data transmission volume, data transmission speed, data source access frequency and number of abnormal interruptions.

[0033] It should be understood that in this embodiment, the adjacent update interval refers to the time span between two adjacent data updates of the data source. For the price data of specialty agricultural products, the market price is constantly changing. The frequency of price changes reflects the frequency of changes in the prices of specialty agricultural products within a certain period of time, and reflects the activity of the agricultural product market and the sensitivity to various internal and external factors. Data transmission volume refers to the amount of data transmitted from the data source to the prediction model within a certain period of time. For the data source of specialty agricultural products, the data transmission volume is a factor to ensure data security and stable acquisition. Compared with the data transmission volume reference value, if the transmission volume is too large, it may exceed the carrying capacity of the storage or network bandwidth, resulting in data loss, transmission interruption and other problems, affecting the integrity and availability of the data; if the transmission volume is too small, it may mean that the data lacks key information and cannot fully reflect various situations related to agricultural product prices. The data transmission speed determines how fast the data reaches the prediction model from the data source. In the price forecast of specialty agricultural products, especially when real-time or timely data acquisition is required for analysis, compared with the data transmission speed reference value, if the speed is too fast, a large amount of data may flow into the data receiving end in a short period of time, causing the data processing system to face the pressure of instantaneous high load, which may cause data caching and untimely processing, resulting in the risk of data loss or incorrect integration; if the speed is too slow, the data cannot reach the prediction model in time, which will cause the prediction results to lag seriously behind the actual market price changes. The data source access frequency indicates the number of access operations to the data source within a specific time period. The number of abnormal interruptions reflects the number of abnormal interruptions that occur during the transmission of the data source. Abnormal interruptions may cause data loss or delays, thereby affecting the accuracy of the prediction.

[0034] The agricultural product data source quality index is obtained based on the agricultural product price forecast data processing, and the preset initial data source quality assessment index threshold is corrected based on the agricultural product data source quality index to obtain the reference data source quality assessment index threshold.

[0035] Specifically, the agricultural product data source quality index is obtained based on the agricultural product price forecast data processing, and the specific analysis process is: obtaining the critical data update frequency index and the critical model expected forecast time range from the agricultural product database; based on the agricultural product price forecast data, a comprehensive analysis is performed to obtain the agricultural product data source quality index, and the agricultural product data source quality index is used to quantify the demand data quality level for price model construction.

[0036] It should be understood that the critical data update frequency index in this embodiment reflects a key frequency indicator for data to maintain sufficient freshness and timeliness in the context of agricultural product price forecasting, and defines the effectiveness boundary of the speed of data update for price model construction. The model's expected prediction time range refers to the time period covered by the model for future predictions of agricultural product prices in the agricultural product price forecasting scenario, and is quantified using natural time units, such as days, weeks, months, quarters, and years. The critical model's expected prediction time range is the time span considered by the price model when predicting agricultural product prices. The number of price changes of highland barley varieties is the number of times agricultural product prices have changed within the expected prediction time range, reflecting the activity level and price fluctuations of the agricultural product market. The reference number of price changes of highland barley varieties is a reference value for the number of price changes of highland barley varieties, which is directly obtained from the agricultural product database.

[0037] In a specific embodiment, the method for obtaining the agricultural product data source quality indicator is:

[0038]

[0039]

[0040] In the formula, G represents the data update frequency index, N i represents the number of price changes of the i-th highland barley variety, represents the average number of price changes of highland barley varieties, X represents the quality index of agricultural product data source, and F b represents the model's expected prediction time range, F0 represents the preset critical model's expected prediction time range, N represents the number of price changes of the reference highland barley variety, ω1 represents the agricultural product data source quality impact weight corresponding to the preset model's expected prediction time range, ω2 represents the agricultural product data source quality impact weight corresponding to the preset data update frequency index, i represents the number of each highland barley variety, i=1, 2, 3, ..., m, m represents the total number of highland barley varieties.

[0041] For example, when ω1 and ω2 are used, the agricultural product data source quality impact weights corresponding to the data update frequency index and the model expected prediction time range can be directly obtained from the agricultural product database. The factors respectively represent the numerical values ​​of the influence of the data update frequency index and the model expected prediction time range on the agricultural product data source quality index. The corresponding relationship can be a pre-set mapping relationship. For example, the types of specialty agricultural products and the agricultural product price impact weights corresponding to the data update frequency index and the model expected prediction time range preset in the agricultural product database form a mapping set. The types of specialty agricultural products are input into the mapping set to obtain the agricultural product data source quality impact weights corresponding to the data update frequency index and the model expected prediction time range. The mapping relationship can be a one-to-one correspondence or a many-to-one relationship. In this embodiment, its value range is (0, 1).

[0042] In this embodiment, the agricultural product data source quality index is used to quantify the quality level of demand data for price model construction. The larger the model's estimated forecast time range, the greater the data update frequency index, the higher the agricultural product data source quality index, and the higher the quality level of demand data for price model construction.

[0043] The algorithm of this embodiment combines the data update frequency index and the model's expected prediction time range, and comprehensively analyzes to obtain the agricultural product data source quality index. In this formula, the data update frequency index and the model's expected prediction time range influence each other. The larger the data update frequency index, the more violent and complex the fluctuations in agricultural product prices in historical periods, which means that when building a price model, a wider time span and more data points are needed to capture and understand the fluctuation pattern. Therefore, the larger the model's expected prediction time range, the higher the requirements for data integrity and representativeness, and the higher the agricultural product data source quality index. By comprehensively analyzing the data update frequency index and the model's expected prediction time range, the agricultural product data source quality index can be accurately obtained, which quantitatively reflects the applicability of the data in the data source to the construction of the price model.

[0044] Specifically, the preset initial data source quality assessment index threshold is corrected according to the agricultural product data source quality index to obtain the reference data source quality assessment index threshold. The specific analysis process is: comparing the agricultural product data source quality index with the data source quality assessment index threshold correction parameters corresponding to each agricultural product data source quality index interval preset in the agricultural product database to obtain the data source quality assessment index threshold correction parameters.

[0045] It should be understood that in this embodiment, the agricultural product database presets the quality index intervals of each agricultural product data source and the data source quality assessment index threshold correction parameters corresponding to each agricultural product data source quality index interval. The agricultural product data source quality index interval is a different range divided according to factors such as the statistical analysis of historical agricultural product price data and the market performance characteristics of different agricultural products. Each quality index interval corresponds to a data source quality assessment index threshold correction parameter. The purpose of setting this correction parameter is to provide corresponding and targeted correction reference values ​​for agricultural product data sources in different quality level intervals to meet the needs of different quality level data sources for initial threshold adjustment. The calculated agricultural product data source quality index is compared with the agricultural product data source quality index interval preset in the database to determine the agricultural product data source quality index interval range, and the corresponding data source quality assessment index threshold correction parameter is obtained accordingly.

[0046] The preset initial data source quality assessment index threshold is summed with the data source quality assessment index threshold correction parameter to obtain a reference data source quality assessment index threshold.

[0047] It should be understood that the initial data source quality assessment index threshold in this embodiment is a relatively general and basic standard value for measuring data source quality that is initially set. It is determined based on historical agricultural product price forecasting experience and common data quality requirements, and is directly obtained from the agricultural product database. The obtained data source quality assessment index threshold correction parameter is summed with the preset initial data source quality assessment index threshold in order to make targeted adjustments to the initial threshold based on the actual situation of the specific agricultural product data source quality indicators, and then obtain the reference data source quality assessment index threshold. The reference data source quality assessment index threshold is a quality assessment standard that truly fits the actual characteristics of current agricultural products and forecasting needs.

[0048] Based on the analysis of agricultural product price forecast data, data source security data and data source validity data, an agricultural product data source evaluation model is established, and the quality evaluation index of each target data source is obtained through processing.

[0049] Specifically, the quality evaluation index of each target data source is obtained by processing, and the specific analysis process is: the data source validity data and the data update frequency index are processed to obtain the validity index of each target data source.

[0050] In a specific embodiment, the target data source effectiveness index is obtained as follows:

[0051]

[0052] Where Y trepresents the effectiveness index of the tth target data source, G represents the data update frequency index, and F t represents the duration of the neighboring update interval of the t-th target data source, ε1 represents the influence weight of the data source effectiveness corresponding to the preset data update frequency index, ε2 represents the influence weight of the data source effectiveness corresponding to the preset neighboring update interval, t represents the number of the target data source, t=1, 2, 3, ..., n, n represents the total number of target data sources.

[0053] For example, when ε1 and ε2 are used, the data source effectiveness corresponding to the data update frequency index and the data source effectiveness influence weight corresponding to the adjacent update interval duration can be directly obtained from the agricultural product database. The factors respectively represent the numerical values ​​of the influence of the data source effectiveness corresponding to the data update frequency index and the adjacent update interval duration on the data source effectiveness index. The corresponding relationship can be a pre-set mapping relationship. For example, the types of characteristic agricultural products and the characteristic agricultural product influence weights corresponding to the data source effectiveness corresponding to the data update frequency index and the adjacent update interval duration obtained from the agricultural product database form a mapping set. The types of characteristic agricultural products are input into the mapping set to obtain the data source effectiveness corresponding to the data update frequency index and the characteristic agricultural product influence weights corresponding to the adjacent update interval duration. The mapping relationship can be a one-to-one correspondence or a many-to-one relationship. In this embodiment, its value range is (0, 1).

[0054] In this embodiment, the data source effectiveness index is used to quantitatively evaluate the ability of the data source to provide effective information. The higher the price fluctuation frequency and the shorter the interval between adjacent updates, the higher the data source effectiveness index and the stronger the ability of the data source to provide effective information.

[0055] The algorithm of this embodiment combines the data source effectiveness and the adjacent update interval duration corresponding to the data update frequency index, and comprehensively analyzes to obtain the data source effectiveness index. In this formula, the data source effectiveness and the adjacent update interval duration corresponding to the data update frequency index influence each other. A high price fluctuation frequency means that agricultural product prices change frequently, while a short adjacent update interval duration indicates that the data source can update the data in a timely manner. When the price fluctuation frequency is higher, the adjacent update interval is shorter, so that the data source can more closely track market price changes, thereby improving the effectiveness of the data source. By comprehensively analyzing the data source effectiveness and the adjacent update interval duration corresponding to the data update frequency index, the data source effectiveness index can be accurately obtained, which quantitatively reflects the ability of the data source to provide effective information.

[0056] The data source security data is processed to obtain the security index of each target data source.

[0057] It should be understood that, in this embodiment, the data transmission volume reference value, the data transmission speed reference value, the data source access frequency critical value and the abnormal interruption number critical value are directly obtained from the database.

[0058] In a specific embodiment, the target data source security index is obtained in the following manner:

[0059]

[0060] In the formula, Z t A represents the security index of the tth target data source, t represents the data transmission volume of the tth target data source, A0 represents the data transmission volume reference value, ΔA represents the preset data transmission volume allowable deviation value, B t represents the data transmission speed of the tth target data source, B0 represents the data transmission speed reference value, ΔB represents the preset data transmission amount allowable deviation value, C t represents the data source access frequency of the tth target data source, C0 represents the data source access frequency critical value, and D t represents the number of abnormal interruptions of the tth target data source, D0 represents the critical value of the number of abnormal interruptions, Indicates the preset data transmission volume impact weight, Indicates the preset data transmission speed impact weight, Indicates the preset data source access frequency impact weight, represents the preset abnormal interruption number impact weight, t represents the number of the target data source, t=1, 2, 3, ..., n, and n represents the total number of the target data sources.

[0061] For example, and When in use, the data source security impact weights corresponding to the data transmission volume, data transmission speed, data source access frequency and number of abnormal interruptions can be directly obtained from the agricultural product database. The factors respectively represent the numerical values ​​of the impact of the data transmission volume, data transmission speed, data source access frequency and number of abnormal interruptions on the data source security index. The corresponding relationship can be a pre-set mapping relationship. For example, the types of characteristic agricultural products and the characteristic agricultural product impact weights corresponding to the data transmission volume, data transmission speed, data source access frequency and number of abnormal interruptions obtained from the agricultural product database form a mapping set. The types of characteristic agricultural products are input into the mapping set to obtain the data source security impact weights corresponding to the data transmission volume, data transmission speed, data source access frequency and number of abnormal interruptions. The mapping relationship can be a one-to-one correspondence or a many-to-one relationship. In this embodiment, its value range is (0, 1).

[0062] In this embodiment, the data source security index is used to quantitatively evaluate the data source's ability to ensure data security. The smaller the deviation between the data transmission volume and the reference value, or the smaller the deviation between the data transmission speed and the reference value, or the lower the data source access frequency, or the fewer abnormal interruptions, the higher the corresponding data source security index, indicating that the data source has a stronger ability to ensure data security.

[0063] The algorithm of this embodiment combines the data transmission volume, data transmission speed, data source access frequency and the number of abnormal interruptions, and comprehensively analyzes to obtain the data source security index. In this formula, the data transmission volume, data transmission speed, data source access frequency and the number of abnormal interruptions influence each other. The larger the data transmission volume, the longer the time required to complete the data transmission, which means the slower the data transmission speed and the more dangerous the target data source. The higher the data source access frequency, the more frequent the data transmission activities, the network congestion or resource competition, which affects the slower data transmission speed and the lower the data security. The more abnormal interruptions, the greater the security risk and the lower the data security. By comprehensively analyzing the data transmission volume, data transmission speed, data source access frequency and the number of abnormal interruptions, the data source security index can be accurately obtained, which quantitatively reflects the level of the data source's ability to ensure data security.

[0064] According to the effectiveness index and security index of each target data source, a comprehensive analysis is performed to obtain the quality evaluation index of each target data source.

[0065] In a specific embodiment, the target data source quality evaluation index is obtained as follows:

[0066]

[0067] Where Y t Represents the effectiveness index of the tth target data source, Z t represents the security index of the tth target data source, β t represents the quality assessment index of the t-th target data source, θ1 represents the preset data source validity data impact weight, θ2 represents the preset data source security data impact weight, t represents the number of the target data source, t=1, 2, 3, ..., n, n represents the total number of target data sources.

[0068] In a specific embodiment, Figure 2 As shown, θ1=θ2=0.5, when Z t =0.2, the functional relationship between the target data source effectiveness index and the target data source quality evaluation index is shown in curve a; when Z t =0.8, the functional relationship between the target data source effectiveness index and the target data source quality evaluation index is shown in curve b; when Z tWhen =1.2, the functional relationship between the target data source effectiveness index and the target data source quality evaluation index is shown as curve c.

[0069] In a specific embodiment, when θ1 and θ2 are used, the target data source quality impact weight corresponding to the data source validity index and the data source security index can be directly obtained from the agricultural product database, and the factors thereof respectively represent the numerical values ​​of the impact of the data source validity index and the data source security index on the target data source quality assessment index, and the corresponding relationship can be a pre-set mapping relationship. For example, the types of characteristic agricultural products and the characteristic agricultural product impact weights corresponding to the data source validity index and the data source security index obtained in the agricultural product database form a mapping set, and the types of characteristic agricultural products are input into the mapping set to obtain the target data source quality impact weight corresponding to the data source validity index and the data source security index. The mapping relationship can be a one-to-one correspondence or a many-to-one relationship. In this embodiment, its value range is (0, 1).

[0070] In this embodiment, the target data source quality evaluation index is used to quantitatively evaluate the overall quality and reliability of each data source in agricultural product price forecasting. The higher the data source validity index and the data source security index, the higher the target data source quality evaluation index, the better the overall quality of the data source in agricultural product price forecasting, and the stronger the reliability.

[0071] The algorithm of this embodiment combines the data source validity index and the data source security index, and comprehensively analyzes to obtain the target data source quality evaluation index. In this formula, the data source validity index and the data source security index influence each other. When the data source validity index is higher and the data source security index is lower, the overall quality evaluation index may be negatively affected, resulting in a decrease in the target data source quality evaluation index. By comprehensively analyzing the data source validity index and the data source security index, the target data source quality evaluation index can be accurately obtained, which quantitatively reflects the adaptability of the data source to the agricultural product price forecasting work.

[0072] The quality assessment index of each target data source is compared with the quality assessment index threshold of the reference data source, and qualified data sources are screened and feedback is provided.

[0073] Specifically, each qualified data source is screened and feedback is provided. The specific analysis process is: the quality assessment index of each target data source is compared with the quality assessment index threshold of the reference data source. If the quality assessment index of a target data source is greater than or equal to the quality assessment index threshold of the reference data source, the data source is marked as a qualified data source; if the quality assessment index of a target data source is less than the quality assessment index threshold of the reference data source, the data source is marked as an unqualified data source; statistics of each qualified data source are obtained and feedback is provided.

[0074] It should be understood that in this embodiment, when the quality assessment index of a target data source is greater than or equal to the reference data source quality assessment index threshold, it means that the data source has reached a preset qualified level in multiple key data quality dimensions, indicating that the data provided by the data source can well meet the needs of related work such as agricultural product price forecasting in terms of accuracy, timeliness and security, and has high quality and high reliability, and can be used as an effective data source for subsequent construction of price forecasting models, data analysis and other operations. When the quality assessment index of a target data source is less than the reference data source quality assessment index threshold, it means that the data source fails to meet the preset standard in terms of overall data quality. It may be that the agricultural product data is not updated in time, resulting in the inability to accurately reflect the latest changes in agricultural product prices; or there are various reasons such as relatively hidden safety hazards in data transmission, affecting the stable use of data, making it unsuitable for work scenarios that require high-quality data such as agricultural product price forecasting. After completing the comparison and marking of all target data sources one by one, the qualified data sources are statistically sorted.

[0075] Specifically, statistics are obtained for each qualified data source and feedback is provided. The specific analysis process is: statistics are performed on the qualified data sources to obtain each qualified data source, original data of each qualified data source is collected, a qualified data source set is constructed, and the qualified data source set is stored in the agricultural product database for constructing an agricultural product price prediction model.

[0076] It should be understood that in this embodiment, the qualified data sources are counted, the number of data sources and basic information that have clearly passed the quality screening are sorted out, and the qualified data sources are listed and key information is sorted out according to the marked results. By collecting the raw data of the qualified data sources, a qualified data source set is constructed, the classified raw data is integrated to form a data whole that is easy to manage and call, and the database technology is used to classify and store it in the agricultural product database by field, so as to achieve long-term data preservation, safe management and convenient call, and help R&D personnel extract data to build models, and help relevant subjects grasp price trends in advance and arrange production and marketing plans reasonably. For example, in the wheat market, statistical screening is performed on qualified data sources such as the planting area and yield data of the agricultural department, the transaction price and volume of the grain trading market, the transportation cost of wheat by logistics companies, and the precipitation and temperature data of the meteorological department. Collect the original data, build a qualified data source set containing wheat-related data of each year and region, and store it in the database. Select a multivariate linear regression model, take wheat yield, transportation cost, market transaction situation and meteorological factors as features, train the model with historical data, and determine the degree of influence of each factor on the price according to the coefficient output by the model, so as to achieve the prediction of wheat prices.

[0077] The agricultural product database is used to store various types of information related to agricultural products, including critical data update frequency index, critical model prediction time range, data transmission volume critical value, data transmission speed critical value, data source access frequency critical value and abnormal interruption number critical value. The data in the agricultural product database can be obtained from channels such as agricultural cooperative information system, agricultural research institution experimental data platform and professional agricultural product information collection and monitoring equipment.

[0078] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that technicians in the relevant technical field can understand and use the present invention well. As long as they do not deviate from the structure of the present invention or exceed the scope defined by the present invention, they should all belong to the protection scope of the present invention.

Claims

1. A method for constructing a price prediction model for characteristic agricultural products, characterized in that: include: Acquire characteristic agricultural product data sources, mark them as target data sources, collect metadata of each target data source, and acquire agricultural product price forecast data, wherein the metadata includes data source validity data and data source security data; The agricultural product data source quality index is obtained according to the agricultural product price forecast data processing, and the preset initial data source quality assessment index threshold is corrected according to the agricultural product data source quality index to obtain the reference data source quality assessment index threshold; Based on the analysis of agricultural product price forecast data, data source security data and data source validity data, an agricultural product data source evaluation model is established, and the quality evaluation index of each target data source is obtained through processing; The quality assessment index of each target data source is compared with the quality assessment index threshold of the reference data source, and qualified data sources are screened and feedback is provided.

2. The method for constructing a characteristic agricultural product price prediction model according to claim 1, characterized in that: The specific analysis process of obtaining agricultural product price forecast data is as follows: Obtain the model's estimated forecast time range and the number of price changes for each highland barley variety; Process the number of price changes of each highland barley variety to obtain the data update frequency index; The data update frequency index and the model's expected forecast time range are jointly labeled as agricultural product price forecast data.

3. The method for constructing a characteristic agricultural product price prediction model according to claim 1, characterized in that: The data source validity data includes the duration of the adjacent update interval; Data source security data includes data transmission volume, data transmission speed, data source access frequency, and number of abnormal interruptions.

4. The method for constructing a characteristic agricultural product price prediction model according to claim 1, characterized in that: The agricultural product data source quality index is obtained by processing the agricultural product price forecast data. The specific analysis process is as follows: Obtain critical data update frequency index and critical model estimated forecast time range from agricultural product database; Based on the agricultural product price forecast data, a comprehensive analysis is conducted to obtain the agricultural product data source quality index, which is used to quantify the quality level of demand data for price model construction.

5. The method for constructing a characteristic agricultural product price prediction model according to claim 1, characterized in that: The preset initial data source quality assessment index threshold is corrected according to the agricultural product data source quality index to obtain the reference data source quality assessment index threshold. The specific analysis process is as follows: Compare the agricultural product data source quality index with the data source quality assessment index threshold correction parameter corresponding to each agricultural product data source quality index interval preset in the agricultural product database to obtain the data source quality assessment index threshold correction parameter; The preset initial data source quality assessment index threshold is summed with the data source quality assessment index threshold correction parameter to obtain a reference data source quality assessment index threshold.

6. The method for constructing a characteristic agricultural product price prediction model according to claim 2, characterized in that: The processing obtains the quality evaluation index of each target data source, and the specific analysis process is as follows: Processing the data source validity data and data update frequency index to obtain the validity index of each target data source; Processing the data source security data to obtain the security index of each target data source; According to the effectiveness index and security index of each target data source, a comprehensive analysis is performed to obtain the quality evaluation index of each target data source.

7. The method for constructing a characteristic agricultural product price prediction model according to claim 1, characterized in that: The screening obtains each qualified data source and provides feedback, and the specific analysis process is as follows: Compare the quality assessment index of each target data source with the quality assessment index threshold of the reference data source. If the quality assessment index of a target data source is greater than or equal to the quality assessment index threshold of the reference data source, mark the data source as a qualified data source. If the quality assessment index of a target data source is less than the quality assessment index threshold of the reference data source, the data source is marked as an unqualified data source; Statistics are obtained for each qualified data source and feedback is provided.

8. The method for constructing a characteristic agricultural product price prediction model according to claim 7, characterized in that: The statistics are obtained from each qualified data source and feedback is given. The specific analysis process is as follows: Qualified data sources are counted to obtain each qualified data source, original data of each qualified data source is collected, a qualified data source set is constructed, and the qualified data source set is stored in the agricultural product database for constructing an agricultural product price prediction model.

9. The method for constructing a characteristic agricultural product price prediction model according to claim 1, characterized in that: The specific analysis process of the quality evaluation index of each target data source is as follows: Where Y t Represents the effectiveness index of the tth target data source, Z t represents the security index of the tth target data source, β t represents the quality assessment index of the t-th target data source, θ1 represents the preset data source validity data impact weight, θ2 represents the preset data source security data impact weight, t represents the number of the target data source, t=1, 2, 3, ..., n, n represents the total number of target data sources.

10. The method for constructing a characteristic agricultural product price prediction model according to claim 6, characterized in that: The effectiveness index of each target data source is used to quantitatively evaluate the practicality of the data in each target data source; The security index of each target data source is used to quantitatively evaluate the reliability of data in each target data source.

Images