A data analysis-based cold-chain logistics data protection method and system

Abstract

The application discloses a cold-chain logistics data protection method and system based on data analysis, relates to the technical field of data analysis, and guarantees the integrity and timeliness of cold-chain transportation data through an original parameter set Par; a characteristic deviation set Dev is generated, which can reveal implicit data deviations in the cold-chain transportation process due to network fluctuations or sensor abnormalities; a data credibility score Qdc is further calculated, so that each batch of cold-chain transportation data can obtain a quantitative credibility result; the data credibility score Qdc is compared with a system threshold value, an abnormality identification set Jud is generated, so that high-credibility data and low-credibility data can be quickly distinguished, and abnormal information is prevented from being directly mixed into a business database; finally, the abnormality identification set Jud is matched with a safety policy set of the cold-chain logistics, targeted control instructions are generated, and it is ensured that abnormal data can trigger immediate protection measures when being identified.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, specifically to a data analysis-based method and system for protecting cold chain logistics data. Background Technology

[0002] Cold chain logistics, as a crucial link in ensuring the quality of fresh food, pharmaceuticals, and vaccines during transportation, has gradually become a core focus in supply chain management and smart logistics in recent years. Driven by information technology and big data, "a data analysis-based method for cold chain logistics data protection" is not merely a technical term; it reflects the higher demands placed on data management in cold chain logistics. Information such as temperature, humidity, energy consumption, location, and time during cold chain transportation needs to be monitored and managed throughout the entire process using data. Therefore, how to improve the integrity and reliability of cold chain logistics data through data analysis without altering hardware conditions has become an important direction for industry development.

[0003] In the current cold chain logistics system, the authenticity and integrity of data often rely on simple verification methods, such as simply checking whether sensors have uploaded data or using single-dimensional temperature and humidity exceedances as the basis for judgment. However, as the complexity of transportation processes increases, these simple verification mechanisms have significant shortcomings: they cannot dynamically reflect the reliability of data during transmission, nor can they accurately identify anomalies caused by delays, missing data, or forgery. This means that once sensor data is interrupted, tampered with, or forged, the cold chain logistics management system cannot make an effective judgment in a timely manner, and may directly incorporate risky data into the decision-making process. Therefore, existing methods have deficiencies in the dynamic scoring of data reliability, and new solutions are urgently needed. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a data protection method and system for cold chain logistics based on data analysis, which solves the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a cold chain logistics data protection method based on data analysis, comprising the following steps: S1. Collect datasets from the cold chain transportation process. The datasets include data integrity rate, collection frequency, and transmission delay, which are combined to form the original parameter set. S2. Extract features from the original parameter set to generate a standardized result set, and perform consistency analysis based on the numerical fluctuations between different collection time points in the same transportation task to construct a self-consistent feature set; then, obtain the feature deviation set by calculating the difference between the standardized result set and the self-consistent feature set. S3. Construct a data credibility score based on the feature deviation set and convert it into a credibility score result; S4. Compare the credibility score with the system's preset credibility threshold, and trigger the generation of an anomaly identifier set based on the comparison results. S5. Match the set of abnormal identifiers with the set of security policies for cold chain logistics, generate corresponding control commands, and issue them for execution.

[0006] Preferably, S1 includes S11; S11. Basic monitoring data during transportation is collected through the vehicle-mounted sensing unit in the cold chain transportation system. The vehicle-mounted sensing unit includes a temperature and humidity sensor, a network communication module, and a local cache module, which can transmit real-time monitoring data to the background data management platform. The data collection method includes two modes: Timed sampling mode: The vehicle-mounted sensing unit triggers data recording once at a preset fixed time interval; Real-time upload mode: When the instantaneous change of any basic monitoring data exceeds the preset threshold, the vehicle-mounted sensing unit triggers data upload, so that abnormal situations are recorded in real time; The basic monitoring data includes data integrity rate Com, acquisition frequency Fre, transmission delay Del, and data identifier ID; The data integrity rate Com is calculated by the ratio of the number of valid data records per unit time to the expected number of data records to be collected. The sampling frequency Fre is obtained by recording the time interval between two consecutive samples and taking the reciprocal. The transmission delay Del is calculated by comparing the difference between the data acquisition time at the sensor and the reception time at the backend.

[0007] Preferably, S1 further includes S12; S12. The acquired basic monitoring data is structured and organized to generate the original parameter set Par; The structured organization is completed through steps S121 and S122; S121. Standardization and organization of data units: The data integrity rate Com was standardized using the maximum-minimum normalization method. The sampling frequency Fre was standardized using a ratio normalization method. The transmission delay Del is standardized using an inverse normalization method. S122. Data Structure Encapsulation and Organization: The normalized data integrity rate Com, acquisition frequency Fre, and transmission delay Del are bound to the corresponding acquisition timestamp and acquisition mode to form the original parameter set Par.

[0008] Preferably, S2 includes S21; S21. Extract features from the parameters in the original parameter set Par to obtain the standardized result set Nor; The standardized result set Nor is specifically extracted and obtained through steps S211 and S212; S211. In the original parameter set Par, for each record, the data integrity rate Com, the acquisition frequency Fre, and the transmission delay Del are calculated using a sliding window to obtain three characteristic indicators within the window: mean, variance, and rate of change. S212. Arrange the three characteristic indicators of mean, variance and rate of change obtained in chronological order, and retain the corresponding timestamps and collection modes to form a standardized result set Nor.

[0009] Preferably, S2 further includes S22; S22. Perform self-consistency analysis on the standardized result set Nor, and use linear regression fitting method to generate the feature deviation set Dev; The self-consistency analysis is specifically completed through steps S221, S222, and S223. S221. In the standardized result set Nor, linear regression modeling is performed using the linear regression fitting method for the time series data of data integrity rate Com, acquisition frequency Fre and transmission delay Del, respectively, to obtain the corresponding regression line equation. S222. Use the regression line equation to calculate the predicted value at each time point, and calculate the difference between the predicted value and the actual observed value at each time point. S223. When the absolute value of the difference exceeds the preset fluctuation threshold, the actual observation value at the current time point, together with the difference size, the corresponding timestamp and parameter category, is recorded in the feature deviation set Dev to identify the actual observation value as a deviation point. The parameter categories include data integrity rate Com, acquisition frequency Fre, and transmission delay Del; After obtaining the feature deviation set Dev, the sum of squared residuals of all differences in the set is compared with the total number of samples to obtain the overall goodness of fit index, which is labeled as the consistency index Con, forming a self-consistent feature set Cns, which is used to represent the consistency level of different parameter categories over the entire time period.

[0010] Preferably, S3 includes S31; S31. Based on the self-consistent feature set Cns, the feature deviation set Dev, and the original parameter set Par, after feature processing, calculate and obtain the data credibility score Qdc. The data credibility score Qdc is obtained using the following formula: Qdc=q1×Com+q2×Fre+q3×(1-Del)+q4×(1-Rat)+q5×Con; In the formula, Com represents the normalized data integrity rate, Fre represents the normalized sampling frequency, Del represents the normalized transmission delay, Rat represents the deviation ratio, Con represents the consistency index, and q1, q2, q3, q4 and q5 represent the weighting coefficients of the normalized data integrity rate Com, the normalized sampling frequency Fre, the normalized transmission delay Del, the deviation ratio Rat and the consistency index Con, respectively, and q1+q2+q3+q4+q5=1, the specific values ​​of which are set by the user. The feature processing is completed through steps S311 and S312; S311. Based on the feature deviation set Dev, the deviation ratio Rat is calculated by the ratio of the number of deviation points to the total number of samples. S312. Based on the consistency index Con, the goodness-of-fit index obtained by the ratio of the sum of squared residuals to the total number of samples is labeled as the self-consistent feature set Cns.

[0011] Preferably, S4 includes S41; S41. Determine the credibility of the data credibility score Qdc and generate the determination result set Jud; The determination result set Jud is specifically obtained through steps S411, S412 and S413. S411. Extract the system's preset confidence threshold Thr, wherein the confidence threshold Thr is limited to the range of [0,1]. S412. Compare the extracted credibility threshold Thr with the data credibility score Qdc to obtain the judgment result; When the data credibility score Qdc is greater than or equal to the credibility threshold Thr, the result is considered credible. When the data credibility score Qdc is less than the credibility threshold Thr, the result is deemed unreliable. S413. Record the judgment result in correspondence with the data identifier ID assigned during data collection, and generate the judgment result set Jud.

[0012] Preferably, S4 further includes S42; S42. Based on the judgment result set Jud, perform decision processing to generate the processing result set Res; The decision-making process is specifically completed through steps S421 and S422; S421. Based on the judgment result set Jud, for data whose judgment result is reliable, it is allowed to enter the data warehouse and subsequent business processes, and is recorded as having been entered into the data warehouse in the processing result set Res. S422. Based on the judgment result set Jud, for data whose judgment result is unreliable, refuse to enter the data warehouse and record it as refused entry into the data warehouse in the processing result set Res. At the same time, trigger the exception marking mechanism and write the data into the exception record database for subsequent analysis.

[0013] Preferably, S5 includes S51; S51. Based on the processing result set Res, perform cold chain logistics data protection operations and issue the execution order, specifically including: For trusted data marked as stored in the result set Res, write it to the trusted storage area and enable data encryption and access control policies. For untrusted data marked as rejected for entry into the database in the processing result set Res, isolate it and store it in an isolated storage area to prevent it from entering the business data flow; For data marked as anomalies in the processing result set Res, write it into the anomaly tracking database, add monitoring flags, and perform special analysis and anomaly tracing manually.

[0014] A data analysis-based cold chain logistics data protection system includes a cold chain data acquisition module, a data extraction module, a deviation extraction module, a reliability assessment module, and a protection decision module. The cold chain data acquisition module collects datasets from the cold chain transportation process. The datasets include data integrity rate, acquisition frequency, and transmission delay, which are combined to form a raw parameter set. The data extraction module extracts features from the original parameter set, generates a standardized result set, and performs consistency analysis based on the numerical fluctuations between different collection time points in the same transportation task to construct a self-consistent feature set; then, by calculating the difference between the standardized result set and the self-consistent feature set, a feature deviation set is obtained. The deviation extraction module constructs a data credibility score based on the feature deviation set and converts it into a credibility score result; The credibility assessment module compares the credibility score with the system's preset credibility threshold and triggers the generation of anomaly identifier set based on the comparison results. The protection decision module matches the set of anomaly identifiers with the set of security policies for cold chain logistics, and generates corresponding control commands for execution.

[0015] This invention provides a data protection method and system for cold chain logistics based on data analysis, which has the following beneficial effects: (1) By collecting data integrity rate Com, collection frequency Fre and transmission delay Del to form the original parameter set Par, the integrity and timeliness of cold chain transportation data are guaranteed. The feature deviation set Dev is generated, which can reveal the hidden data deviation caused by network fluctuations or sensor anomalies during cold chain transportation. The data credibility score Qdc is further calculated so that each batch of cold chain transportation data can obtain a quantitative credibility result. By comparing the data credibility score Qdc with the system threshold, the anomaly identifier set Jud is generated, which can quickly distinguish between high-credibility data and low-credibility data and avoid directly mixing abnormal information into the business database. Finally, the anomaly identifier set Jud is matched with the security policy set of cold chain logistics to generate targeted control instructions, which ensures that the abnormal data can trigger immediate protection measures when it is identified.

[0016] (2) By performing sliding window calculations on the data integrity rate Com, acquisition frequency Fre, and transmission delay Del in the original parameter set Par, three characteristic indicators—mean, variance, and rate of change—are obtained and combined to form a standardized result set Nor. This not only filters out interference from occasional fluctuations but also reveals situations where the surface mean is normal but the internal structure is unstable. A data trend model is established using linear regression fitting, and the characteristic deviation set Dev and the self-consistency characteristic set Cns are obtained through residual analysis. This enables the precise location of local anomalies and the global assessment of the overall trend, avoiding the serious situation of complete data upload blockage. This ensures early warning and location of potential anomalies in cold chain transportation, preventing the quality deterioration of medicines or fresh food during transportation due to delayed discovery.

[0017] (3) By integrating the self-consistency feature set Cns, the feature deviation set Dev, and the original parameter set Par, the data reliability score Qdc is calculated, which can reflect the overall reliability of the data with a unified value. By comparing the data reliability score Qdc with the reliability threshold Thr set by the system, a judgment result set Jud is generated, and a processing result set Res is further formed, so that the final destination of each piece of data is accurately marked. Based on the processing result set Res, the data is classified into a trusted storage area, an isolated storage area, or an anomaly tracking library, and differentiated protection is achieved by combining encryption and access control strategies. This not only ensures the secure isolation and reliable storage of cold chain transportation data, but also provides a traceability basis when data anomalies occur, avoiding decision-making errors in the pharmaceutical or fresh food delivery process caused by the spread of single-point errors to the business layer. Attached Figure Description

[0018] Figure 1 This is a schematic diagram illustrating the steps of a cold chain logistics data protection method based on data analysis according to the present invention. Figure 2This is a schematic diagram of a cold chain logistics data protection system based on data analysis according to the present invention. Figure 3 This is a schematic diagram illustrating the trend of the consistency index Con for the self-consistent feature set Cns. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0020] Example 1 This invention provides a data protection method for cold chain logistics based on data analysis. Please refer to [link / reference]. Figure 1 This includes the following steps: S1. Collect datasets from the cold chain transportation process. The datasets include data integrity rate, collection frequency, and transmission delay, which are combined to form the original parameter set. S2. Extract features from the original parameter set to generate a standardized result set, and perform consistency analysis based on the numerical fluctuations between different collection time points in the same transportation task to construct a self-consistent feature set; then, obtain the feature deviation set by calculating the difference between the standardized result set and the self-consistent feature set. S3. Construct a data credibility score based on the feature deviation set and convert it into a credibility score result; S4. Compare the credibility score with the system's preset credibility threshold, and trigger the generation of an anomaly identifier set based on the comparison results. S5. Match the set of abnormal identifiers with the set of security policies for cold chain logistics, generate corresponding control commands, and issue them for execution.

[0021] In this embodiment, by forming the original parameter set Par in step S1 using the data integrity rate Com, the acquisition frequency Fre, and the transmission delay Del, the integrity and timeliness of the cold chain transportation data are guaranteed. In step S2, a self-consistent feature set Cns is constructed through standardization and consistency analysis, and a feature deviation set Dev is generated, which can reveal the implicit data deviations caused by network fluctuations or sensor anomalies during cold chain transportation. In step S3, the data credibility score Qdc is further calculated, so that each batch of cold chain transportation data can obtain a quantitative credibility result, instead of the traditional method. The previous method relied solely on the presence or absence of a single parameter for a rough assessment. In step S4, an anomaly identifier set, Jud, is generated by comparing the data reliability score Qdc with the system threshold. This allows for rapid differentiation between high-reliability and low-reliability data, preventing the direct mixing of abnormal information into the business database. Finally, in step S5, the anomaly identifier set Jud is matched with the cold chain logistics safety policy set to generate targeted control commands. For example, when the data reliability score Qdc is too low, the system immediately issues control commands such as "keep the freezer closed" or "activate the backup sensor," thus preventing serious consequences of cold chain disruptions caused by data errors. This approach not only improves the authenticity and reliability of cold chain logistics data but also ensures that anomaly data triggers immediate protective measures upon identification, avoiding the risk of spoilage of medicines or fresh food during transportation due to inaccurate temperature monitoring data.

[0022] Example 2 Specifically: S1 includes S11; S11. Basic monitoring data during transportation is collected through the vehicle-mounted sensing unit in the cold chain transportation system. The vehicle-mounted sensing unit includes a temperature and humidity sensor, a network communication module, and a local cache module, which can transmit real-time monitoring data to the background data management platform. The data collection method includes two modes: Timed sampling mode: The vehicle-mounted sensing unit triggers data recording once at a preset fixed time interval, such as once every 30 seconds. In this mode, the uniformity and comparability of the data can be guaranteed. Real-time upload mode: When the instantaneous change of any basic monitoring data exceeds the preset threshold (e.g., temperature change exceeds 0.5℃ or network latency exceeds 500 milliseconds), the vehicle-mounted sensing unit triggers data upload, so that abnormal situations are recorded in real time; The switching strategy between the two modes is as follows: the system defaults to running in timed sampling mode. When any parameter triggers the threshold condition, it automatically switches to real-time upload mode. After the abnormal situation subsides and the system returns to stability, it switches back to timed sampling mode. In this way, it can ensure data recording efficiency under normal circumstances and respond quickly in case of emergencies. The basic monitoring data includes data integrity rate Com, acquisition frequency Fre, transmission delay Del, and data identifier ID; The data integrity rate Com is calculated by the ratio of the number of valid data records per unit time to the expected number of data records to be collected. The sampling frequency Fre is obtained by recording the time interval between two consecutive samples and taking the reciprocal. The transmission delay Del is calculated by comparing the difference between the data acquisition time at the sensor and the reception time at the backend. It should be noted that: The vehicle-mounted sensing unit not only has data acquisition capabilities, but also ensures data upload through a network communication module; in the event of network failure, the local caching module ensures that data is not lost, thereby indirectly improving data integrity. The reason for setting up two modes is that relying solely on the timed sampling mode may result in insufficient data response, while relying solely on the real-time upload mode will cause unnecessary network resource consumption during the stable phase; therefore, by combining the two, the overall efficiency and robustness of the system can be improved. The mode switching logic uses threshold conditions to ensure that communication resources are saved in a stable environment and that data is timely during sudden fluctuations. This mechanism not only improves the rationality of the acquisition frequency Fre and transmission delay Del, but also enhances the reliable monitoring effect of the entire cold chain transportation process.

[0023] S1 further includes S12; S12. The acquired basic monitoring data is structured and organized to generate the original parameter set Par; The structured organization is completed through steps S121 and S122; S121. Standardization and organization of data units: The data integrity rate Com is standardized using the min-max normalization method, and the formula is: (actual value) (Minimum value) ÷ (Maximum value) The minimum value is 0%, where the maximum value is 100% and the minimum value is 0%; the normalized result is mapped to the interval [0,1]. The sampling frequency Fre is standardized using a ratio normalization method, which is the ratio of the actual sampling frequency to the preset reference sampling frequency. When the ratio is greater than 1, it is counted as 1, and when the ratio is less than 0, it is counted as 0, to ensure that the normalization result is in the range of [0,1]. The transmission delay Del is standardized using an inverse normalization method, i.e., 1 (Actual delay ÷ Reference delay threshold), when the delay exceeds the threshold, it is counted as 0, and when the delay is 0, it is counted as 1. This is to ensure that the lower the delay value, the higher the normalization result. S122. Data Structure Encapsulation and Organization: The normalized data integrity rate Com, sampling frequency Fre, and transmission delay Del are bound to the corresponding sampling timestamp and sampling mode (timed sampling mode or real-time upload mode) to form the original parameter set Par; Each record in the original parameter set Par contains a normalized parameter value, a timestamp, and a collection mode to ensure consistency and traceability in subsequent processing.

[0024] In this embodiment, by setting a timed sampling mode and a real-time upload mode in the vehicle-mounted sensing unit and dynamically switching them based on threshold logic, not only is the sampling uniformity ensured during normal transportation, but real-time upload is also triggered immediately in case of sudden temperature changes or network latency spikes. This avoids the problems of insufficient data response or resource waste caused by a single acquisition mode. For example, when a pharmaceutical transport vehicle suddenly experiences signal attenuation on a mountain road, the transmission delay Del may instantly exceed 500 milliseconds. In this case, the real-time upload mode will ensure that the anomaly is uploaded and recorded in a timely manner, allowing managers to detect the risk within seconds. At the same time, the local caching module can save data when there is a network anomaly, avoiding a decrease in data integrity Com due to communication interruption, thereby ensuring the data availability of the entire transportation task in extreme scenarios. Furthermore, after data unit standardization and data structure encapsulation, the generated original parameter set Par not only unifies the data representation of different dimensions, but also binds a timestamp and acquisition mode to each record, enabling rapid tracing of data source and context in subsequent analysis. It significantly improves the reasonable control of the acquisition frequency Fre and transmission delay Del in cold chain transportation, realizes continuous monitoring and anomaly protection of key parameters such as temperature and humidity in actual complex transportation environments, and ultimately avoids monitoring blind spots caused by short-term fluctuations or communication loss.

[0025] Example 3 Please see Figure 1 and Figure 3 Specifically: S2 includes S21; S21. Extract features from the parameters in the original parameter set Par to obtain the standardized result set Nor; The standardized result set Nor is specifically extracted and obtained through steps S211 and S212; S211. In the original parameter set Par, for each record, the data integrity rate Com, the acquisition frequency Fre, and the transmission delay Del are calculated using a sliding window to obtain three characteristic indicators within the window: mean, variance, and rate of change. S212. Arrange the three characteristic indicators of mean, variance and rate of change obtained in chronological order, and retain the corresponding timestamps and collection modes to form a standardized result set Nor. It should be noted that: The role of the mean: The mean can reflect the overall level of data within a time window. For example, if the mean of the sampling frequency Fre is too low, it often indicates that the monitoring device has a problem with untimely data collection; if the mean of the transmission delay Del is too high, it indicates that the network transmission performance has been poor for a long time. If only single-point data is used, occasional fluctuations may be misjudged as abnormalities, while the mean can filter out short-term interference. The role of variance: Variance is used to characterize the stability of data within a window. For example, if the data completeness rate Com fluctuates drastically at different time points, although the mean may seem normal, the variance will be amplified, revealing the unreliability of data continuity. If the variance index is missing, it is easy to miss the situation of "normal mean but internal instability". The role of the rate of change: The rate of change reflects the trend of data. For example, a gradual increase in transmission delay Del may mean that the network link is about to become congested, but the mean and variance may not be significant in the early stage. If the rate of change indicator is missing, the response to the gradually deteriorating trend will be slow, and the early warning time will be delayed.

[0026] S2 further includes S22; S22. Perform self-consistency analysis on the standardized result set Nor, and use linear regression fitting method to generate the feature deviation set Dev; The self-consistency analysis is specifically completed through steps S221, S222, and S223. S221. In the standardized result set Nor, linear regression modeling is performed using the linear regression fitting method for the time series data of data integrity rate Com, acquisition frequency Fre and transmission delay Del, respectively, to obtain the corresponding regression line equation. S222. Use the regression line equation to calculate the predicted value at each time point, and calculate the difference between the predicted value and the actual observed value at each time point. S223. When the absolute value of the difference exceeds the preset fluctuation threshold, the actual observation value at the current time point, together with the difference size, the corresponding timestamp and parameter category, is recorded in the feature deviation set Dev to identify the actual observation value as a deviation point. The parameter categories include data integrity rate Com, acquisition frequency Fre, and transmission delay Del; After obtaining the feature deviation set Dev, the sum of squared residuals of all differences in the set is compared with the total number of samples to obtain the overall goodness of fit index, which is marked as the consistency index Con, forming a self-consistent feature set Cns, which is used to represent the consistency level of different parameter categories over the entire time period. After obtaining the consistency index Con, the consistency index Con is compared with the preset consistency threshold Thr2. When the consistency index Con < the consistency threshold Thr2, it is determined that the corresponding parameter category has a long-term deviation risk, and the risk is marked in the self-consistent feature set Cns. When the consistency index Con≥the consistency thresholdThr2, it is determined that there is no long-term deviation risk in the corresponding parameter category, and it is marked as risk-free in the self-consistent feature set Cns; This approach enables the early detection of trend deterioration, avoiding delayed responses due to reliance solely on data reliability scores (Qdc). It should be noted that: Linear regression fitting method: The regression line equation is established by using the least squares method, so that the variation pattern of cold chain data under stable conditions can be clearly modeled, and local outliers are revealed by residual analysis. The feature deviation set Dev consists of observations that exceed the fluctuation threshold. The set includes the actual observation, the magnitude of the difference, the timestamp, and the parameter category. The purpose of this set is to clearly identify observations that are abnormal under normal trends, ensuring that abnormal data can be accurately located and classified. The self-consistency feature set Cns is a goodness-of-fit index calculated by the ratio of the sum of squared residuals to the total number of samples. It is used to measure the overall consistency level of the entire parameter sequence. If the self-consistency feature set Cns shows too low consistency, it indicates that there is a long-term deviation or systematic anomaly in the cold chain data.

[0027] In this embodiment, by performing sliding window calculations on the data integrity rate Com, sampling frequency Fre, and transmission delay Del in the original parameter set Par, three characteristic indicators—mean, variance, and rate of change—are obtained and combined to form a standardized result set Nor. This not only filters out interference from occasional fluctuations but also reveals situations where the surface mean is normal but the internal stability is not. For example, during a cold chain vehicle's nighttime operation, although the mean of the sampling frequency Fre remains within the normal range, the variance index remains consistently high, indicating intermittent interruptions in the actual sensor sampling. If only the mean is relied upon for judgment, this potential problem may be missed. In step S22, a data trend model is established using a linear regression fitting method, and the characteristic deviation set Dev and the self-consistency feature set Cns are obtained through residual analysis. This enables both accurate local anomaly location and global trend assessment. For example, when the transmission delay Del shows a gradual increasing trend within 30 minutes, even if the initial mean and variance are not obviously abnormal, the self-consistency feature set Cns will show a decrease in consistency, thus providing an early warning of potential network link congestion and avoiding a severe situation of complete data upload blockage. This ensures early warning and location of potential anomalies in cold chain transportation, preventing the quality of medicines or fresh food from deteriorating during transportation due to delayed detection.

[0028] Example 4 Specifically: S3 includes S31; S31. Based on the self-consistent feature set Cns, the feature deviation set Dev, and the original parameter set Par, after feature processing, calculate and obtain the data credibility score Qdc. The data credibility score Qdc is obtained using the following formula: Qdc=q1×Com+q2×Fre+q3×(1-Del)+q4×(1-Rat)+q5×Con; In the formula, Com represents the normalized data integrity rate, Fre represents the normalized sampling frequency, Del represents the normalized transmission delay, Rat represents the deviation ratio, Con represents the consistency index, and q1, q2, q3, q4 and q5 represent the weighting coefficients of the normalized data integrity rate Com, the normalized sampling frequency Fre, the normalized transmission delay Del, the deviation ratio Rat and the consistency index Con, respectively, and q1+q2+q3+q4+q5=1, the specific values ​​of which are set by the user. The feature processing is completed through steps S311 and S312; S311. Based on the feature deviation set Dev, the deviation ratio Rat is calculated by the ratio of the number of deviation points to the total number of samples. S312. Based on the consistency index Con, the goodness-of-fit index obtained by the ratio of the sum of squared residuals to the total number of samples is labeled as the self-consistent feature set Cns. It should be noted that: The feature deviation set Dev is a set of records made during the consistency analysis of cold chain transportation data when the difference between the predicted value and the actual observed value exceeds a preset fluctuation threshold. Content: Each record includes the actual observed value, the magnitude of the difference, the timestamp, and the parameter category; Function: Used to clearly mark abnormal observation points that occur under normal trends, so as to achieve accurate location and classification of local deviations; The consistency index Con is a single index obtained by calculating the ratio of the sum of squared residuals of all differences in the characteristic deviation set Dev to the total number of samples. Value characteristics: The value ranges from 0 to 1. The larger the value, the more stable and consistent the trend of the parameter category is over the entire time series. Purpose: Used to quantitatively measure the overall consistency level of a single parameter category (e.g., data integrity rate Com), helping to determine the stability of this parameter during cold chain transportation; The self-consistency feature set Cns is a set composed of consistency indices Con corresponding to different parameter categories (including data integrity rate Com, acquisition frequency Fre, and transmission delay Del). Content: Includes the consistency index Con value for each parameter category; Function: It is used to comprehensively reflect the overall consistency level of cold chain transportation data across multiple dimensions, thereby providing a basis for comprehensive judgment.

[0029] S4 includes S41; S41. Determine the credibility of the data credibility score Qdc and generate the determination result set Jud; The determination result set Jud is specifically obtained through steps S411, S412 and S413. S411. Extract the system's preset confidence threshold Thr, wherein the confidence threshold Thr is limited to the range of [0,1] and is used to distinguish whether the data is qualified. S412. Compare the extracted credibility threshold Thr with the data credibility score Qdc to obtain the judgment result; When the data credibility score Qdc is greater than or equal to the credibility threshold Thr, the result is considered credible. When the data credibility score Qdc is less than the credibility threshold Thr, the result is deemed unreliable. S413. Record the judgment result in correspondence with the data identifier ID assigned during data collection, and generate the judgment result set Jud.

[0030] S4 also includes S42; S42. Based on the judgment result set Jud, perform decision processing to generate the processing result set Res; The decision-making process is specifically completed through steps S421 and S422; S421. Based on the judgment result set Jud, for data whose judgment result is reliable, it is allowed to enter the data warehouse and subsequent business processes, and is recorded as having been entered into the data warehouse in the processing result set Res. S422. Based on the judgment result set Jud, for data whose judgment result is unreliable, refuse to enter the data warehouse and record it as refused entry into the data warehouse in the processing result set Res. At the same time, trigger the anomaly marking mechanism to write the data into the anomaly record database for subsequent analysis. It should be noted that: Processing Result Set Res: Records the final processing status of all data, including data identifier ID, judgment label, and processing action (data already entered into the database / data entry rejected / abnormal marker). Anomaly marking mechanism: Untrusted data is recorded in an anomaly log database instead of being discarded directly, so as to facilitate subsequent anomaly tracing and improved analysis.

[0031] S5 includes S51; S51. Based on the processing result set Res, perform cold chain logistics data protection operations and issue the execution order, specifically including: For trusted data marked as stored in the result set Res, write it to the trusted storage area and enable data encryption and access control policies to ensure its confidentiality and integrity during transmission and storage. For untrusted data marked as rejected for entry into the database in the processing result set Res, it is isolated and stored in the isolation storage area to prevent it from entering the business data flow, while retaining a copy of the data for subsequent traceability; For data marked as anomalies in the processing result set Res, write it into the anomaly tracking database, add monitoring tags, and perform special analysis and anomaly tracing manually; It should be noted that: Trusted Storage Area: A secure area specifically for storing trusted data, equipped with encrypted storage and access control mechanisms; Isolated storage area: Untrusted data is stored in isolation to avoid contaminating the core data warehouse, while retaining traceability capabilities.

[0032] In this embodiment, by fusing the self-consistent feature set Cns, the feature deviation set Dev, and the original parameter set Par, a data credibility score Qdc is calculated, which can reflect the overall reliability of the data with a unified value. In step S4, by comparing the data credibility score Qdc with the credibility threshold Thr set by the system, a judgment result set Jud is generated, and a processing result set Res is further formed, so that the final destination of each piece of data is accurately marked. In step S5, based on the processing result set Res, the data is classified into a trusted storage area, an isolated storage area, or an anomaly tracking library, and differentiated protection is achieved by combining encryption and access control strategies. For example, in a vaccine cold chain transportation mission, if the data credibility score Qdc is lower than the credibility threshold Thr, the data will be recorded as "untrustworthy" and automatically isolated and stored in an isolated storage area to prevent it from entering the business database and causing subsequent vaccine inventory statistics errors. Trustworthy data, on the other hand, is encrypted and written to a trusted storage area to ensure it is not tampered with or leaked during cross-network transmission and storage. Simultaneously, data marked as "anomaly" is written to an anomaly tracking database for maintenance personnel to conduct source tracing analysis in conjunction with vehicle transportation information. This not only ensures the secure isolation and reliable storage of cold chain transportation data but also provides traceability evidence in the event of data anomalies, preventing single-point errors from spreading to the business layer and causing decision-making errors in the pharmaceutical or fresh food distribution process.

[0033] Example 5 A data analysis-based cold chain logistics data protection system, please refer to... Figure 2 Specifically, it includes a cold chain data acquisition module, a data extraction module, a deviation extraction module, a reliability assessment module, and a protection decision module; The cold chain data acquisition module collects datasets from the cold chain transportation process. The datasets include data integrity rate, acquisition frequency, and transmission delay, which are combined to form a raw parameter set. The data extraction module extracts features from the original parameter set, generates a standardized result set, and performs consistency analysis based on the numerical fluctuations between different collection time points in the same transportation task to construct a self-consistent feature set; then, by calculating the difference between the standardized result set and the self-consistent feature set, a feature deviation set is obtained. The deviation extraction module constructs a data credibility score based on the feature deviation set and converts it into a credibility score result; The credibility assessment module compares the credibility score with the system's preset credibility threshold and triggers the generation of anomaly identifier set based on the comparison results. The protection decision module matches the set of anomaly identifiers with the set of security policies for cold chain logistics, and generates corresponding control commands for execution.

[0034] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A data protection method for cold chain logistics based on data analysis, characterized in that: Includes the following steps: S1. Collect datasets from the cold chain transportation process. The datasets include data integrity rate, collection frequency, and transmission delay, which are combined to form the original parameter set. S2. Extract features from the original parameter set to generate a standardized result set, and perform consistency analysis based on the numerical fluctuations between different collection time points in the same transportation task to construct a self-consistent feature set; then, obtain the feature deviation set by calculating the difference between the standardized result set and the self-consistent feature set. S3. Construct a data credibility score based on the feature deviation set and convert it into a credibility score result; S4. Compare the credibility score with the system's preset credibility threshold, and trigger the generation of an anomaly identifier set based on the comparison results. S5. Match the set of abnormal identifiers with the set of security policies for cold chain logistics, generate corresponding control commands, and issue them for execution.

2. The cold chain logistics data protection method based on data analysis according to claim 1, characterized in that: S1 includes S11; S11. Basic monitoring data during transportation is collected through the vehicle-mounted sensing unit in the cold chain transportation system. The vehicle-mounted sensing unit includes a temperature and humidity sensor, a network communication module, and a local cache module, which can transmit real-time monitoring data to the background data management platform. The data collection method includes two modes: Timed sampling mode: The vehicle-mounted sensing unit triggers data recording once at a preset fixed time interval; Real-time upload mode: When the instantaneous change of any basic monitoring data exceeds the preset threshold, the vehicle-mounted sensing unit triggers data upload, so that abnormal situations are recorded in real time; The basic monitoring data includes data integrity rate Com, acquisition frequency Fre, transmission delay Del, and data identifier ID; The data integrity rate Com is calculated by the ratio of the number of valid data records per unit time to the expected number of data records to be collected. The sampling frequency Fre is obtained by recording the time interval between two consecutive samples and taking the reciprocal. The transmission delay Del is calculated by comparing the difference between the data acquisition time at the sensor and the reception time at the backend.

3. The cold chain logistics data protection method based on data analysis according to claim 2, characterized in that: S1 further includes S12; S12. The acquired basic monitoring data is structured and organized to generate the original parameter set Par; The structured organization is completed through steps S121 and S122; S121. Standardization and organization of data units: The data integrity rate Com was standardized using the maximum-minimum normalization method. The sampling frequency Fre was standardized using a ratio normalization method. The transmission delay Del is standardized using an inverse normalization method. S122. Data Structure Encapsulation and Organization: The normalized data integrity rate Com, acquisition frequency Fre, and transmission delay Del are bound to the corresponding acquisition timestamp and acquisition mode to form the original parameter set Par.

4. The cold chain logistics data protection method based on data analysis according to claim 3, characterized in that: S2 includes S21; S21. Extract features from the parameters in the original parameter set Par to obtain the standardized result set Nor; The standardized result set Nor is specifically extracted and obtained through steps S211 and S212; S211. In the original parameter set Par, for each record, the data integrity rate Com, the acquisition frequency Fre, and the transmission delay Del are calculated using a sliding window to obtain three characteristic indicators within the window: mean, variance, and rate of change. S212. Arrange the three characteristic indicators of mean, variance and rate of change obtained in chronological order, and retain the corresponding timestamps and collection modes to form a standardized result set Nor.

5. A cold chain logistics data protection method based on data analysis according to claim 4, characterized in that: S2 further includes S22; S22. Perform self-consistency analysis on the standardized result set Nor, and use linear regression fitting method to generate the feature deviation set Dev; The self-consistency analysis is specifically completed through steps S221, S222, and S223. S221. In the standardized result set Nor, linear regression modeling is performed using the linear regression fitting method for the time series data of data integrity rate Com, acquisition frequency Fre and transmission delay Del, respectively, to obtain the corresponding regression line equation. S222. Use the regression line equation to calculate the predicted value at each time point, and calculate the difference between the predicted value and the actual observed value at each time point. S223. When the absolute value of the difference exceeds the preset fluctuation threshold, the actual observed value at the current time point, along with the difference size, the corresponding timestamp, and the parameter category, is recorded in the feature deviation set Dev. The parameter categories include data integrity rate Com, acquisition frequency Fre, and transmission delay Del; After obtaining the feature deviation set Dev, the sum of squared residuals of all differences in the set is compared with the total number of samples to obtain the overall goodness of fit index, which is marked as the consistency index Con, forming the self-consistent feature set Cns.

6. The cold chain logistics data protection method based on data analysis according to claim 5, characterized in that: S3 includes S31; S31. Based on the self-consistent feature set Cns, the feature deviation set Dev, and the original parameter set Par, after feature processing, calculate and obtain the data credibility score Qdc. The data credibility score Qdc is obtained using the following formula: Qdc=q1×Com+q2×Fre+q3×(1-Del)+q4×(1-Rat)+q5×Con; In the formula, Com represents the normalized data integrity rate, Fre represents the normalized sampling frequency, Del represents the normalized transmission delay, Rat represents the deviation ratio, Con represents the consistency index, and q1, q2, q3, q4 and q5 represent the weighting coefficients of the normalized data integrity rate Com, the normalized sampling frequency Fre, the normalized transmission delay Del, the deviation ratio Rat and the consistency index Con, respectively, and q1+q2+q3+q4+q5=1, the specific values ​​of which are set by the user. The feature processing is completed through steps S311 and S312; S311. Based on the feature deviation set Dev, the deviation ratio Rat is calculated by the ratio of the number of deviation points to the total number of samples. S312. Based on the consistency index Con, the goodness-of-fit index obtained by the ratio of the sum of squared residuals to the total number of samples is labeled as the self-consistent feature set Cns.

7. A cold chain logistics data protection method based on data analysis according to claim 6, characterized in that: S4 includes S41; S41. Determine the credibility of the data credibility score Qdc and generate the determination result set Jud; The determination result set Jud is specifically obtained through steps S411, S412 and S413. S411. Extract the system's preset confidence threshold Thr, wherein the confidence threshold Thr is limited to the range of [0,1]. S412. Compare the extracted credibility threshold Thr with the data credibility score Qdc to obtain the judgment result; When the data credibility score Qdc is greater than or equal to the credibility threshold Thr, the result is considered credible. When the data credibility score Qdc is less than the credibility threshold Thr, the result is deemed unreliable. S413. Record the judgment result in correspondence with the data identifier ID assigned during data collection, and generate the judgment result set Jud.

8. A cold chain logistics data protection method based on data analysis according to claim 7, characterized in that: S4 also includes S42; S42. Based on the judgment result set Jud, perform decision processing to generate the processing result set Res; The decision-making process is specifically completed through steps S421 and S422; S421. Based on the judgment result set Jud, for data whose judgment result is reliable, it is allowed to enter the data warehouse and subsequent business processes, and is recorded as having been entered into the data warehouse in the processing result set Res. S422. Based on the judgment result set Jud, for data whose judgment result is unreliable, refuse to enter the data warehouse and record it as refused entry into the data warehouse in the processing result set Res. At the same time, trigger the exception marking mechanism and write the data into the exception record database for subsequent analysis.

9. A cold chain logistics data protection method based on data analysis according to claim 8, characterized in that: S5 includes S51; S51. Based on the processing result set Res, perform cold chain logistics data protection operations and issue the execution order, specifically including: For trusted data marked as stored in the result set Res, write it to the trusted storage area and enable data encryption and access control policies. For untrusted data marked as rejected for entry into the database in the processing result set Res, isolate it and store it in an isolated storage area to prevent it from entering the business data flow; For data marked as anomalies in the processing result set Res, write it into the anomaly tracking database, add monitoring flags, and perform special analysis and anomaly tracing manually.

10. A data analysis-based cold chain logistics data protection system, applied to the data analysis-based cold chain logistics data protection method described in any one of claims 1 to 9, characterized in that: It includes a cold chain data acquisition module, a data extraction module, a deviation extraction module, a reliability assessment module, and a protection decision module; The cold chain data acquisition module collects datasets from the cold chain transportation process. The datasets include data integrity rate, acquisition frequency, and transmission delay, which are combined to form a raw parameter set. The data extraction module extracts features from the original parameter set, generates a standardized result set, and performs consistency analysis based on the numerical fluctuations between different collection time points in the same transportation task to construct a self-consistent feature set; then, by calculating the difference between the standardized result set and the self-consistent feature set, a feature deviation set is obtained. The deviation extraction module constructs a data credibility score based on the feature deviation set and converts it into a credibility score result; The credibility assessment module compares the credibility score with the system's preset credibility threshold and triggers the generation of anomaly identifier set based on the comparison results. The protection decision module matches the set of anomaly identifiers with the set of security policies for cold chain logistics, and generates corresponding control commands for execution.

Images