Reinforcing plate production quality data processing method based on big data

Abstract

The invention relates to the field of data processing, in particular to a stiffening plate production quality data processing method based on big data, and the method comprises the steps: obtaining the production quality data of a stiffening plate, and calculating the distribution uniformity of all data in the production quality data; any item of data in the production quality data is used as target item data, the ratio of the distribution uniformity degree of the target item data to the average value of the distribution uniformity degrees of all items of data is calculated and multiplied by the initial segmentation segment number, the segmentation segment number of the target item data is obtained, the optimal segmentation ratio is determined based on the segmentation segment number, and the eccentric optimization degree is calculated; and taking the eccentric optimization degree as a segmentation optimization degree to carry out isolated tree division on each item of data to obtain a division result, constructing an isolated forest according to the division result, and carrying out anomaly detection on the data points. According to the method, the optimal segmentation ratio and the segmentation optimization degree are dynamically calculated, so that different data distribution and characteristics can be adapted, and the accuracy of data processing is improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and more specifically, to a method for processing reinforcing plate production quality data based on big data. Background Art

[0002] In modern manufacturing, reinforcing plates are used in fields such as construction, transportation, and machinery to enhance the stability and bearing capacity of structures. With the advancement of industry and intelligent manufacturing, a large amount of data is generated in the production process, including raw material quality, production environment parameters, equipment operation status, real-time monitoring data, and quality inspection results. The accumulation of these data provides a valuable source of information for improving the production quality of reinforcing plates. However, traditional quality control methods often rely on manual inspection and experience judgment, and have problems such as low efficiency, slow response, and insufficient accuracy. Therefore, there is an urgent need for a reinforcing plate production quality data processing method based on big data, so that enterprises can monitor the production process in real time, identify anomalies in a timely manner, and predict product quality, thereby achieving precise control and optimization.

[0003] The existing Chinese patent application document with the publication number CN118643444A discloses a big data anomaly detection, device, equipment, storage medium and computer program product. The method improves the original isolation forest algorithm. First, multiple isolated trees are created according to the business data to be processed; then, the fitness value of each isolated tree in the multiple isolated trees is calculated; then, the fitness value of the isolated tree is used as a screening index, and the target isolated tree is selected from the multiple isolated trees based on the taboo search algorithm to form an isolated forest. Finally, the business data to be processed is detected for anomalies according to the isolated forest to obtain anomaly detection results. The fitness value of the isolated tree is used as a screening index, and the isolated trees with lower fitness values ​​and more redundant are removed from the multiple isolated trees based on the taboo search algorithm, and the better target isolated trees are screened out, which not only reduces the space occupied by the isolated forest, but also reduces the computational overhead of anomaly detection and improves the efficiency of anomaly detection.

[0004] The application document uses a taboo search algorithm to remove isolated trees with low fitness values ​​and redundancy from multiple isolated trees, thereby solving the problem that constructing more isolated trees consumes a lot of memory space, has a large computational overhead, and has low anomaly detection efficiency. However, in the process of anomaly detection of reinforcement plate production quality data by the isolation forest algorithm, the conventional data division by random thresholds cannot take into account the degree of splitting of each data in the reinforcement plate production quality data, resulting in inaccurate threshold division for each data, and normal data is mistakenly divided into abnormal data, resulting in inaccurate processing of reinforcement plate production quality data. Summary of the invention

[0005] In order to solve the problem that when the isolation forest algorithm performs anomaly detection on the reinforcement plate production quality data, the random threshold is used to divide the data and cannot take into account the degree of splitting of each data in the reinforcement plate production quality data, resulting in inaccurate data processing, the present invention provides solutions in the following aspects.

[0006] The method for processing production quality data of reinforcing plates based on big data comprises: obtaining production quality data of reinforcing plates, calculating the distribution uniformity of each data in the production quality data; taking any data in the production quality data as target item data, calculating the ratio between the distribution uniformity of the target item data and the average value of the distribution uniformity of all item data, and multiplying it by the initial segmentation number to obtain the segmentation number of the target item data, determining the optimal segmentation ratio based on the segmentation number, calculating the eccentricity optimization degree, and taking the eccentricity optimization degree as the segmentation optimization degree; performing isolation tree partitioning on each data according to the segmentation optimization degree to obtain the partitioning result, constructing an isolation forest according to the partitioning result, and performing anomaly detection on the data points; wherein the eccentricity optimization degree satisfies the following relationship: , where Indicates The data point in In the data The degree of eccentricity preference during the sub-segmentation, Indicates The data point in Item data The maximum absolute value of the difference between the maximum values ​​of the item data, Indicates Item data The optimal split ratio for splitting the data. Indicates The data point in Item data The minimum absolute value of the difference between the minimum values ​​of the data items.

[0007] The effect is as follows: calculating the distribution uniformity of each data item can more accurately evaluate the discreteness and distribution characteristics of the data, and determining the number of segmentation segments according to the ratio of the distribution uniformity to the average distribution uniformity, so that the segmentation is more adapted to the actual distribution of the data, thereby improving the accuracy of the segmentation. Determining the optimal segmentation ratio helps to more effectively identify abnormal data points in the process of building an isolation forest, and replaces random threshold division with adaptive threshold division, reducing the risk of misjudging normal data as abnormal data.

[0008] Preferably, the production quality data includes: steel strength, steel thickness, equipment pressure, production environment temperature, humidity, defective product rate and production efficiency.

[0009] Preferably, the distribution uniformity of each data includes: Sort the target item data from small to large, take the absolute difference between two adjacent data points in the target item data and the difference between the average range of the data points in the target item data as the fluctuation deviation, and normalize them; The average value of the normalized sum of the fluctuation deviations is used as the distribution uniformity of the target item data.

[0010] The effect is that the average value of the normalized sum of fluctuation deviations can reflect the fluctuation and dispersion in the data set, and help identify the distribution characteristics of the data, such as whether it is skewed, concentrated or dispersed. Since abnormal data points often show large deviations in distribution uniformity, abnormal data points can be identified more accurately.

[0011] Preferably, the distribution uniformity of the various data also includes: The target item data are sorted from small to large, and any data point in the target item data is used as the benchmark data. The average value of the sum of the relative absolute differences between the benchmark data and the target item data is calculated to obtain the distribution uniformity of the target item data.

[0012] The effect is that by calculating the relative absolute difference between each data point and the mean, the degree of deviation of the data point from the central trend of the entire data set can be quantified, thereby evaluating the uniformity of the data distribution.

[0013] Preferably, the optimal split ratio includes: The target item data is segmented based on the number of segmentation segments, and the optimal segmentation ratio of the remaining data in the target item data after each segmentation is calculated; The remaining data in the target item data is taken as the data to be segmented, and the inverse of the difference between the number of segmented segments and the current segmentation index plus 1 is taken as the optimal segmentation ratio of the data to be segmented.

[0014] The effect is that as each segmentation proceeds, the number of remaining data points decreases, and the optimal segmentation ratio is dynamically adjusted to ensure that the data is distributed as evenly as possible at each segmentation, thereby improving segmentation efficiency and uniformity. Dynamic adjustment based on the actual distribution of the data enhances the algorithm's adaptability to different data distributions.

[0015] Preferably, the process of dividing the isolated tree is: Obtain the segmentation preference of each data point in the target item data, use the maximum segmentation preference as the threshold, and perform the first data division; After the first data division, the data with the larger range in the two divided parts is taken as the data segment to be divided. Then, the segmentation preference of each data point in the data segment to be divided is calculated, and the threshold selection and data division are continued until the target item data is divided the number of times to reach the number of segmentation segments minus 1. The second segmentation is constructed according to the division result. An isolated tree of item data.

[0016] The effect is that by selecting the data point with the maximum segmentation preference as the threshold for segmentation, potential abnormal data points can be identified more accurately because these points have a higher degree of distinction in the data set. By giving priority to the data points with the highest segmentation preference, the data can be segmented into representative data segments more quickly, thereby improving the overall segmentation efficiency.

[0017] Preferably, an isolation forest is constructed according to the partitioning results, and anomaly detection is performed on the data points, including: An isolation forest is constructed based on the isolation trees of each data item, and the abnormality degree of each data point is obtained based on the layer depth of the data point in each isolation tree. Data points whose abnormality degree is greater than an abnormality degree threshold are marked as abnormal data points.

[0018] Preferably, the segmentation preference degree further includes: Taking any data point in the target item data as the benchmark data, calculate the difference between the average value of the data points in the target item data that is greater than the benchmark data and the average value of the data points that is less than the benchmark data, and the square of the ratio between the difference and the range of the data segment to be segmented in the target item data, and obtain the segmentation difference degree; The product of the eccentricity preference degree and the segmentation difference degree is taken as the segmentation preference degree.

[0019] Preferably, the segmentation preference degree further includes: Taking any data point in the target item data as the benchmark data, the difference between the median of the data points of the target item data to be segmented that is larger than the benchmark data and the median of the data points that is smaller than the benchmark data, and the ratio between the median of the data points and the range of the data segment to be segmented in the target item data are taken as the segmentation difference degree; The product of the eccentricity preference degree and the segmentation difference degree is taken as the segmentation preference degree.

[0020] The present invention has the following effects: 1. The present invention can adapt to different data distributions and characteristics by dynamically calculating the optimal segmentation ratio and segmentation preference degree, making data processing more flexible. It does not rely on a fixed threshold, but determines the segmentation strategy based on the actual distribution of the data, thereby enhancing the algorithm's adaptability to different data sets and different production conditions, and enhancing the accuracy of segmentation.

[0021] 2. The present invention can more accurately determine the data segmentation point by calculating the distribution uniformity and eccentricity optimization degree of each data, thereby more effectively identifying abnormal data points when constructing an isolated tree, improving the efficiency of anomaly detection, and being able to effectively distinguish normal data from abnormal data, thereby reducing false positives and negatives, and improving the accuracy and sensitivity of anomaly detection.

[0022] 3. The present invention can more effectively utilize computing resources, reduce unnecessary calculations, and improve data processing efficiency by giving priority to processing data segments with larger ranges and selecting data points with the maximum segmentation preference as thresholds. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 It is a method flow chart of steps S1 to S3 in the method for processing quality data of reinforcement plate production based on big data in an embodiment of the present invention. DETAILED DESCRIPTION

[0024] 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 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 those skilled in the art without creative work are within the scope of protection of the present invention.

[0025] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0026] Reference Figure 1 The reinforcement plate production quality data processing method based on big data includes steps S1 to S3, which are as follows: S1: Obtain the production quality data of the reinforcement plate and calculate the distribution uniformity of each item in the production quality data.

[0027] The production quality data that needs to be collected include but are not limited to: steel strength, steel thickness, equipment pressure, production environment temperature, humidity, defective product rate and production efficiency.

[0028] It should be noted that each data is collected during the production process of the reinforcement plates. Through big data, the production quality data of multiple batches of reinforcement plates are collected. For each batch of reinforcement plate production quality data, one data is collected for each item. For a batch of reinforcement plates, it is considered that the production quality of the entire batch is the same, so only one data point is collected.

[0029] Further analysis shows that the conventional isolation forest algorithm divides each data item separately through random thresholds, and then obtains the abnormality of a data point according to the layer depth of a data point in all data items. However, random threshold division often fails to achieve the effect of reasonable data division, and normal data and abnormal data are divided into the same category, resulting in inaccurate calculation of data abnormality. Therefore, the uniformity of distribution of each data item is obtained, including: Sort the target item data from small to large, take the absolute difference between two adjacent data points in the target item data and the difference between the average range of the data points in the target item data as the fluctuation deviation, and normalize them; The average value of the normalized sum of the fluctuation deviations is used as the distribution uniformity of the target item data.

[0030] Specifically, the distribution uniformity satisfies the following relationship: ; In the formula, Indicates The uniformity of the distribution of the data items, Represents the number of data points for each item of data, Indicates Item data data, Indicates Item data data, Indicates The data is extremely poor.

[0031] That is to say, for any data, the more uniform the data distribution is, the less it needs to be segmented, and the more uneven the data is, the more it should be segmented.

[0032] In addition, another embodiment further includes: The target item data are sorted from small to large, and any data point in the target item data is used as the benchmark data. The average value of the sum of the relative absolute differences between the benchmark data and the target item data is calculated to obtain the distribution uniformity of the target item data.

[0033] Specifically, the distribution uniformity satisfies the following relationship: ; In the formula, Indicates The uniformity of the distribution of the data items, Represents the number of data points for each item of data, Indicates Item data data, Indicates The average value in the data items.

[0034] That is to say, It indicates the degree to which each data point deviates from the mean, and can be used to assess the dispersion of the data, that is, the dispersion of each data point around the mean.

[0035] Further analysis shows that for any data, the segmentation point should be obtained closer to the data boundary. For unevenly distributed data, the more times it should be segmented, the less data will be divided out each time to achieve a division fineness that matches the number of segmentation segments.

[0036] S2: Take any item of production quality data as the target item data, calculate the ratio between the distribution uniformity of the target item data and the average value of the distribution uniformity of all item data, and multiply it by the initial number of segmentation segments to obtain the number of segmentation segments of the target item data, determine the optimal segmentation ratio based on the number of segmentation segments, calculate the eccentricity preference degree, and use the eccentricity preference degree as the segmentation preference degree.

[0037] In this embodiment, the number of initial segmentation segments is , the implementers can set the initial number of segments according to the actual implementation situation.

[0038] Specifically, the number of segmentation segments satisfies the following relationship: ; In the formula, Indicates The number of segments of the item data, Indicates the number of initial segmentation segments, Indicates The uniformity of the distribution of the data items, It represents the average value of the distribution uniformity of all item data. Represents the ceiling function.

[0039] In other words, the distribution uniformity of each data item affects the number of segments, and then the initial number of segments is adjusted. The greater the distribution uniformity, the more uniform the distribution of the data is, and the number of segments can be appropriately reduced. Conversely, the smaller the distribution uniformity, the more uneven the distribution of the data is, and the number of segments needs to be increased in order to divide the data more finely.

[0040] Optimal split ratio, including: Taking any item of production quality data as target item data, segmenting the target item data based on the number of segmentation segments, and calculating the optimal segmentation ratio of the remaining data in the target item data after each segmentation; The remaining data in the target item data is taken as the data to be segmented, and the inverse of the difference between the number of segmented segments and the current segmentation index plus 1 is taken as the optimal segmentation ratio of the data to be segmented.

[0041] It should be noted that in After the split, the number of segments is However, the remaining number of segments does not include the segment currently being considered, i.e. Therefore, we need to add 1 to include the current segment. By adding 1, we ensure that at each split, we take into account the number of remaining segments including the current split.

[0042] Specifically, the optimal split ratio satisfies the following relationship: ; In the formula, Indicates Item data The optimal percentage of the split, Indicates The number of segments of the item data, Indicates the segment number index.

[0043] That is to say, taking a data item as an example, get the number of segments of the data item In an ideal state, each segment of the data is evenly divided, so the result of the first division is , which means splitting from all data As the first segmentation data, each segmentation will reduce the amount of remaining data to be segmented. When the data is split for the second time, the remaining amount of data to be split is the initial amount of data minus the previous amount. After each split, the amount of data remaining to be split decreases, so the ratio of each subsequent split is It increases relative to the remaining data to ensure that all data is eventually segmented, where The maximum value of , that is, A total of 100 items of data need to be processed. Second division.

[0044] In addition, another embodiment further includes: The quartiles of the data points are used to determine the split points, and the proportion of data points in the data segment to be split is set to 25% to ensure that each split interval contains a similar proportion of data points.

[0045] The first embodiment of obtaining the degree of eccentricity preference: The eccentricity optimization degree satisfies the following relationship: ; In the formula, Indicates The data point in In the data The degree of eccentricity preference during the sub-segmentation, Indicates The data point in Item data The maximum absolute value of the difference between the maximum values ​​of the item data, Indicates Item data The optimal split ratio for splitting the data. Indicates The data point in Item data The minimum absolute value of the difference between the minimum values ​​of the data items.

[0046] That is to say, for the The division ratio is expected to be In each partitioning process, the partitioned data is compared with the first The degree of eccentricity preference of each data is obtained by multiplying the distance between the two extreme values ​​of the data to be divided by the proportion of all data. By adding an optimal percentage value to the minimum value in the proportion, the closer the data proportion is to the optimal percentage, the higher the degree of eccentricity preference is.

[0047] It should be noted that in the first embodiment, the degree of eccentricity preference is calculated by combining the relative positions of the data points to the maximum and minimum values ​​and the optimal segmentation ratio. This method helps to determine the best data segmentation point so that the segmented data segments are more uniform, while taking into account the relative positions and priorities of the data points.

[0048] In addition, in the second embodiment, the segmentation preference degree also includes: Taking any data point in the target item data as the benchmark data, calculate the difference between the average value of the data points in the target item data that is greater than the benchmark data and the average value of the data points that is less than the benchmark data, and the square of the ratio between the difference and the range of the data segment to be segmented in the target item data, and obtain the segmentation difference degree; The product of the eccentricity preference degree and the segmentation difference degree is taken as the segmentation preference degree.

[0049] Specifically, the segmentation preference degree satisfies the following relationship: ; In the formula, Indicates The data point in The preferred degree of segmentation in the item data, Indicates The data point in In the data The degree of eccentricity preference during the sub-segmentation, Indicates Item data The data segments to be split are larger than The average value of all data of data, Indicates Item data The data segments to be split are smaller than The average value of all data of data, Indicates Item data The range of the data segments to be segmented.

[0050] It should be noted that, in the second embodiment, based on the first embodiment, the calculation of the segmentation difference degree is added. When any data is divided, it is not only necessary to divide it proportionally, but also the division should be able to make the difference between the two divided data as large as possible. Therefore, it is necessary to comprehensively consider the segmentation difference degree of each data. The segmentation preference is calculated by combining the eccentric preference of the data point and the position difference relative to other data points. It helps to determine the best data segmentation point, making the segmented data segments more uniform. The segmentation difference calculation in the formula uses a square operation to make larger differences more significant, which helps to identify the segmentation point, so that the segmented data segments are more uniform.

[0051] In addition, in the third embodiment, the segmentation preference degree also includes: Taking any data point in the target item data as the benchmark data, the difference between the median of the data points of the target item data to be segmented that is larger than the benchmark data and the median of the data points that is smaller than the benchmark data, and the ratio between the median of the data points and the range of the data segment to be segmented in the target item data are taken as the segmentation difference degree; The product of the eccentricity preference degree and the segmentation difference degree is taken as the segmentation preference degree.

[0052] Specifically, the segmentation preference degree satisfies the following relationship: ; In the formula, Indicates The data point in Item data, Indicates The data point in In the data The degree of eccentricity preference during the sub-segmentation, Indicates Item data The median of the data segments to be segmented is greater than the target item data. Indicates Item data The data segments to be segmented are smaller than the median of the target item data. Indicates Item data The range of the data segments to be segmented.

[0053] It should be noted that the third embodiment is similar in logic to the second embodiment, and the difference lies in the way of calculating the degree of segmentation difference. By using the median and range, this method can better capture the distribution characteristics of the data, especially when the data distribution is asymmetric or has extreme values, so as to make the segmented data segments more uniform.

[0054] S3: Perform isolation tree partitioning on each data point according to the degree of eccentricity optimization to obtain the partitioning result, construct an isolation forest based on the partitioning result, and perform anomaly detection on the data points.

[0055] The process of isolation tree partitioning is: Obtain the segmentation preference of each data point in the target item data, use the maximum segmentation preference as the threshold, and perform the first data division; After the first data division, the data with the larger range in the two divided parts is taken as the data segment to be divided. Then, the segmentation preference of each data point in the data segment to be divided is calculated, and the threshold selection and data division are continued until the target item data is divided the number of times to reach the number of segmentation segments minus 1. The second segmentation is constructed according to the division result. An isolated tree of item data.

[0056] An isolation forest is constructed based on the isolated trees of each data item. The abnormality degree of each data point is obtained based on the layer depth of the data point in each isolated tree, and the data points with abnormality degrees greater than the abnormality degree threshold are marked as abnormal data points.

[0057] In this embodiment, the abnormality threshold is 0.9, and the implementer can set the abnormality threshold according to the actual situation. Further, the batch of reinforcement plates corresponding to the abnormal data points can be checked, the cause of the data abnormality can be analyzed, and the production and construction management can be strengthened.

[0058] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0059] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A method for processing production quality data of reinforcing plates based on big data, characterized in that: include: Obtaining production quality data of the reinforcement plate, and calculating the distribution uniformity of each item of the production quality data; Taking any item of the production quality data as the target item data, calculating the ratio between the distribution uniformity of the target item data and the average value of the distribution uniformity of all item data, and multiplying it by the initial segmentation number to obtain the segmentation number of the target item data, determining the optimal segmentation ratio based on the segmentation number, calculating the eccentricity preference degree, and taking the eccentricity preference degree as the segmentation preference degree; According to the segmentation optimization degree, each data is divided into isolated trees to obtain the division results. Based on the division results, an isolation forest is constructed and anomaly detection is performed on the data points. The eccentricity optimization degree satisfies the following relationship: , where Indicates The data point in In the data The degree of eccentricity preference during the sub-segmentation, Indicates The data point in Item data The maximum absolute value of the difference between the maximum values ​​of the item data, Indicates Item data The optimal split ratio for the second data split, Indicates The data point in Item data The minimum absolute value of the difference between the minimum values ​​of the data items.

2. The method for processing reinforcing plate production quality data based on big data according to claim 1, characterized in that: The production quality data include: steel strength, steel thickness, equipment pressure, production environment temperature, humidity, defective product rate and production efficiency.

3. The method for processing reinforcing plate production quality data based on big data according to claim 1, characterized in that: The distribution uniformity of the data includes: Sort the target item data from small to large, take the absolute difference between two adjacent data points in the target item data and the difference between the average range of the data points in the target item data as the fluctuation deviation, and normalize them; The average value of the normalized sum of the fluctuation deviations is used as the distribution uniformity of the target item data.

4. The method for processing reinforcing plate production quality data based on big data according to claim 1, characterized in that: The distribution uniformity of the data also includes: The target item data are sorted from small to large, and any data point in the target item data is used as the benchmark data. The average value of the sum of the relative absolute differences between the benchmark data and the target item data is calculated to obtain the distribution uniformity of the target item data.

5. The method for processing reinforcing plate production quality data based on big data according to claim 1, characterized in that: The optimal split ratio includes: The target item data is segmented based on the number of segmentation segments, and the optimal segmentation ratio of the remaining data in the target item data after each segmentation is calculated; The remaining data in the target item data is taken as the data to be segmented, and the inverse of the difference between the number of segmented segments and the current segmentation index plus 1 is taken as the optimal segmentation ratio of the data to be segmented.

6. The method for processing reinforcing plate production quality data based on big data according to claim 1, characterized in that: The process of isolated tree partitioning is as follows: Obtain the segmentation preference of each data point in the target item data, use the maximum segmentation preference as the threshold, and perform the first data division; After the first data division, the data with the larger range in the two divided parts is taken as the data segment to be divided. Then, the segmentation preference of each data point in the data segment to be divided is calculated, and the threshold selection and data division are continued until the target item data is divided the number of times to reach the number of segmentation segments minus 1. The second segmentation is constructed according to the division result. An isolated tree of item data.

7. The method for processing reinforcing plate production quality data based on big data according to claim 1, characterized in that: Based on the partition results, an isolation forest is constructed and anomaly detection is performed on data points, including: An isolation forest is constructed based on the isolation trees of each data item, and the abnormality degree of each data point is obtained based on the layer depth of the data point in each isolation tree. Data points whose abnormality degree is greater than an abnormality degree threshold are marked as abnormal data points.

8. The method for processing reinforcing plate production quality data based on big data according to claim 1, characterized in that: The segmentation preference degree also includes: Taking any data point in the target item data as the benchmark data, calculate the difference between the average value of the data points in the target item data that is greater than the benchmark data and the average value of the data points that is less than the benchmark data, and the square of the ratio between the difference and the range of the data segment to be segmented in the target item data, and obtain the segmentation difference degree; The product of the eccentricity preference degree and the segmentation difference degree is taken as the segmentation preference degree.

9. The method for processing reinforcing plate production quality data based on big data according to claim 1, characterized in that: The segmentation preference degree also includes: Taking any data point in the target item data as the benchmark data, the difference between the median of the data points of the target item data to be segmented that is larger than the benchmark data and the median of the data points that is smaller than the benchmark data, and the ratio between the median of the data points and the range of the data segment to be segmented in the target item data are taken as the segmentation difference degree; The product of the eccentricity preference degree and the segmentation difference degree is taken as the segmentation preference degree.

Images