A data processing method for physical examination ERP management system

Abstract

The embodiment of the present invention discloses a data processing method for a physical examination ERP management system, and relates to the field of data processing. The method includes: obtaining the visual examination data of employees and establishing a sample space; mapping the visual examination data to the sample space to obtain multiple sample space data; clustering the sample space data using the ISODATA algorithm to obtain multiple result clusters; calculating the data aggregation degree corresponding to each result cluster according to the data distribution characteristics of the result cluster, and merging adjacent result clusters at least according to the similarity of the data aggregation degree to obtain a preferred cluster; determining the target compression method corresponding to each preferred cluster based on the data aggregation degree and compression loss of the preferred cluster, wherein the target compression method includes lossy compression and lossless compression; compressing and storing the visual examination data in the corresponding preferred cluster according to the target compression method.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to a data processing method for a physical examination ERP management system. Background Art

[0002] ERP (Enterprise Resource Planning) management system is an integrated information system that helps organizations effectively manage their business processes. The core goal of ERP system is to improve the efficiency and effectiveness of the organization. In ERP management system, physical examination data can be seen as a component specific to employee health and safety management. This data usually falls under the scope of human resource management function, focusing on tracking and managing employee health status, physical examination results and related medical records.

[0003] Currently, when employees' vision examination data are stored through the ERP management system, the vision data is generally stored based on the employee's name. However, there are a large number of employees in an enterprise, and the vision examination data of multiple employees may be similar. If the above-mentioned corresponding storage method is used, a large amount of redundant space may be generated, which is not conducive to improving the efficiency and coordination of the ERP system.

[0004] Based on this, it is necessary to study a more scientific data processing method for physical examination ERP management system to improve the storage efficiency of ERP management system. Summary of the invention

[0005] To solve the above problems, the present specification provides a data processing method for a physical examination ERP management system, the method comprising:

[0006] Obtain employees' vision examination data and establish sample space;

[0007] Mapping the vision examination data to the sample space to obtain a plurality of sample space data;

[0008] Clustering the sample space data using the ISODATA algorithm to obtain multiple result clusters;

[0009] Calculating the data aggregation degree corresponding to each result cluster according to the data distribution characteristics of the result cluster, and merging adjacent result clusters at least according to the similarity of the data aggregation degrees to obtain a preferred cluster;

[0010] Determining a target compression method corresponding to each preferred cluster based on the data concentration and compression loss of the preferred cluster, wherein the target compression method includes lossy compression and lossless compression;

[0011] The vision examination data in the corresponding preferred cluster is compressed and stored according to the target compression method.

[0012] In some embodiments, calculating the data aggregation degree corresponding to each result cluster according to the data distribution characteristics of the result cluster includes:

[0013] Process the result clusters using PCA decomposition to obtain multiple principal component vectors, and store them in sequence A;

[0014] Determining the morphological characteristics of the result cluster according to the modulus of the largest principal component vector in the sequence A and the average of the modulus of all principal component vectors;

[0015] Calculating the average density of the data in the result cluster according to the amount of data in the result cluster and the convex hull volume of the result cluster;

[0016] The data aggregation degree corresponding to the result cluster is determined based on the morphological features, the average density of the data, the average distance between the data in the result cluster and the cluster center position, and the average distance between the result cluster and adjacent clusters.

[0017] In some embodiments, determining the data aggregation degree corresponding to the result cluster based on the morphological features, the average density of the data, the average distance between the data in the result cluster and the cluster center position, and the average distance between the result cluster and adjacent clusters includes:

[0018] The difference between the modulus of the largest principal component vector in the sequence A and the average value of the modulus of all principal component vectors is used as the morphological feature of the result cluster;

[0019] Multiplying the morphological feature, the average data density, and the average distance between the data in the result cluster and the cluster center position to obtain a first product;

[0020] The ratio of the average distance between the result cluster and the adjacent clusters to the first product is used as an index, and an exponential function is used for mapping to obtain the data aggregation degree corresponding to the result cluster.

[0021] In some embodiments, merging adjacent result clusters at least according to the similarity of the data aggregation to obtain a preferred cluster includes:

[0022] For any target result cluster a;

[0023] Determine the similarity between the target result cluster a and the adjacent result cluster b according to the difference between the data aggregation degree corresponding to the target result cluster a and the data aggregation degree corresponding to the adjacent result cluster b;

[0024] Determine the relative data size of the adjacent result cluster b according to the data size of the adjacent result cluster b and the maximum and minimum values ​​of the data sizes of all result clusters;

[0025] Based on the Euclidean distance between the cluster center positions of the target result cluster a and the adjacent result cluster b, the similarity, and the relative data size, calculating the possibility of merging the target result cluster a and the adjacent result cluster b;

[0026] The target result cluster a is merged with one or more adjacent result clusters b according to a first preset threshold and the merging possibility to obtain the preferred cluster.

[0027] In some embodiments, the calculating the possibility of merging the target result cluster a and the adjacent result cluster b based on the Euclidean distance between the cluster center positions of the target result cluster a and the adjacent result cluster b, the similarity, and the relative data size includes:

[0028] Multiplying the Euclidean distance between the cluster center positions of the target result cluster a and the adjacent result cluster b, the similarity, and the relative data volume to obtain a second product;

[0029] The second product is normalized to obtain a possibility of merging the target result cluster a with the adjacent result cluster b.

[0030] In some embodiments, determining a target compression method corresponding to each preferred cluster based on the data concentration and compression loss of the preferred cluster includes:

[0031] For each target preferred cluster;

[0032] Calculating the data aggregation degree corresponding to the target preferred cluster;

[0033] Determining the relative data aggregation degree of the target preferred cluster according to the data aggregation degree corresponding to the target preferred cluster and the maximum value of the data aggregation degrees corresponding to all preferred clusters;

[0034] Calculate the compression coefficient corresponding to the target preferred cluster according to the relative data concentration and the cumulative sum of the difference between each data in the target preferred cluster and the average value of the data in the cluster;

[0035] A target compression mode corresponding to the target preferred cluster is determined according to the compression coefficient.

[0036] In some embodiments, the calculating the data aggregation degree corresponding to the preferred cluster includes:

[0037] Calculate the average data aggregation degree corresponding to all result clusters merged by the target preferred cluster;

[0038] The average data aggregation degree is used as the data aggregation degree corresponding to the target preferred cluster.

[0039] In some embodiments, the calculating the compression coefficient corresponding to the target preferred cluster according to the relative data concentration and the cumulative sum of the difference between each data in the target preferred cluster and the average value of the data in the cluster includes:

[0040] Multiplying the relative data concentration by the cumulative sum of the differences between each data in the target preferred cluster and the average value of the data in the cluster to obtain a third product;

[0041] The third product is used as the compression coefficient corresponding to the target preferred cluster.

[0042] In some embodiments, determining the target compression mode corresponding to the target preferred cluster according to the compression coefficient includes:

[0043] Normalizing the compression coefficient to obtain a normalized result;

[0044] When the normalized result corresponding to the target preferred cluster is less than or equal to a second preset threshold, determining the target compression mode corresponding to the target preferred cluster as lossy compression;

[0045] When the normalized result corresponding to the target preferred cluster is greater than the second preset threshold, the target compression mode corresponding to the target preferred cluster is determined to be lossless compression.

[0046] In some embodiments, the first preset threshold is 0.25, and the second preset threshold is 0.3.

[0047] The data processing method of the physical examination ERP management system provided in the embodiments of this specification may bring at least the following beneficial effects:

[0048] (1) By clustering and merging the sample space data corresponding to the employee vision examination data, one or more preferred clusters are obtained. Then, different compression storage methods are determined by calculating the aggregation degree and compression loss of the preferred clusters. This can improve the redundant space caused by direct storage, save storage space, and improve the storage efficiency of the ERP management system.

[0049] (2) By determining the relative data aggregation of the target preferred cluster and the maximum value of the data aggregation of all preferred clusters, and then calculating the compression coefficient of the target preferred cluster according to the relative data aggregation and the cumulative sum of the difference between each data in the target preferred cluster and the average value of the data in the cluster, not only can the aggregation and compression loss of the preferred cluster be considered at the same time, but also the compression loss corresponding to the target preferred cluster when using lossy compression can be amplified by calculating the cumulative sum, making its loss characteristics more obvious, thereby improving the accuracy of the target compression method determined based on this;

[0050] (3) By normalizing the compression coefficient corresponding to the target preferred cluster, and then when the corresponding normalized result is less than or equal to the second preset threshold, the corresponding target compression method is determined as lossy compression, and when the corresponding normalized result is greater than the second preset threshold, the corresponding target compression method is determined as lossless compression, thereby saving storage space while ensuring the accuracy and reliability of the data.

[0051] Additional features will be described in part in the following description. This specification will become apparent to those skilled in the art by reviewing the following and accompanying drawings, or may be understood by the production or operation of the examples. The features of this specification may be implemented and obtained by practicing or using various aspects of the methods, tools, and combinations described in the following detailed examples. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] This specification will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0053] Figure 1 is an exemplary flow chart of a data processing method of a physical examination ERP management system according to some embodiments of this specification;

[0054] Figure 2 It is a schematic diagram of sample space data mapping according to some embodiments of this specification. DETAILED DESCRIPTION

[0055] In order to more clearly illustrate the technical solutions of the embodiments of this specification, the following is a brief introduction to the drawings required for the description of the embodiments. Obviously, the drawings described below are only some examples or embodiments of this specification. For ordinary technicians in this field, this specification can also be applied to other similar scenarios based on these drawings without creative work. Unless it is obvious from the language environment or otherwise explained, the same reference numerals in the figures represent the same structure or operation.

[0056] Flowcharts are used in this specification to illustrate the operations performed by the system according to the embodiments of this specification. It should be understood that the preceding or following operations are not necessarily performed precisely in order. Instead, the steps may be processed in reverse order or simultaneously. At the same time, other operations may also be added to these processes, or one or more operations may be removed from these processes.

[0057] The data processing method of the physical examination ERP management system provided by the embodiment of this specification is described in detail below with reference to the accompanying drawings.

[0058] Figure 1is an exemplary flow chart of a data processing method of a physical examination ERP management system according to some embodiments of this specification. In some embodiments, the data processing method 100 of the physical examination ERP management system can be executed by processing logic, which can include hardware (e.g., circuits, dedicated logic, programmable logic, microcode, etc.), software (instructions running on a processing device to perform hardware simulation), etc., or any combination thereof. In some embodiments, Figure 1 One or more operations in the flowchart of the data processing method 100 of the physical examination ERP management system can be implemented by a processing device and / or a terminal device. For example, the data processing method 100 of the physical examination ERP management system can be stored in a storage device in the form of instructions and called and / or executed by a processing device and / or a terminal device.

[0059] Reference Figure 1 In some embodiments, the data processing method 100 of the physical examination ERP management system may include:

[0060] S110, obtaining the employee's vision examination data and establishing a sample space.

[0061] In some embodiments, the employee's vision examination data can be collected through the employee's vision examination results. Specifically, in some embodiments of this specification, the vision examination data can refer to refractive examination data. The embodiments of this specification can collect the refractive examination data of the employee's eyes and perform subsequent processing.

[0062] Refractive examination data is usually recorded in the following main parameters: sphere, cylinder and axis. Among them, sphere indicates the myopia or hyperopia of the eye; cylinder indicates the need for astigmatism correction; axis is used for astigmatism correction, which indicates the direction of astigmatism and the angle at which the cylinder correction needs to be placed, ranging from 0 to 180 degrees.

[0063] In some embodiments, the vision examination data of each employee may be distinguished and recorded by left and right eyes respectively. For example, in some embodiments, the vision examination data of an employee may be represented as follows:

[0064] Right eye (OD): -1.25 (spherical) -0.50 (cylindrical) x 90° (axial)

[0065] Left eye (OS): -1.00 (spherical) -0.75 (cylindrical) x 85° (axial)

[0066] Furthermore, in order to better process the employee's vision examination data in the subsequent process, a corresponding sample space can be established based on the characteristic dimension of the vision examination data to be used for converting the vision examination data. Figure 2As shown, in some embodiments, the sample space can be established with the spherical lens, cylindrical lens and axial position as a dimension respectively, so as to obtain a three-dimensional sample space.

[0067] S120, mapping the vision examination data to the sample space to obtain a plurality of sample space data.

[0068] Continue to refer to Figure 2 , the employee's vision examination data can be distinguished by left and right eyes, and mapped to the above sample space respectively, so as to obtain a data point respectively (refer to Figure 2 Each data point is a sample space data.

[0069] In the embodiment of the present specification, the vision examination data of all employees can be mapped to the above sample space, so as to obtain a plurality of sample space data, wherein each sample space data can be represented in the form of a three-dimensional vector.

[0070] S130, clustering the sample space data using the ISODATA algorithm to obtain a plurality of result clusters.

[0071] Since the data processing method provided in the embodiments of this specification needs to be compressed and stored according to the similarity of the sample space data, and the similarity data is manifested as the clustering characteristics of data points in the three-dimensional sample space, the clustered data can be stored using lossy compression according to the size of its value, thereby saving storage space. For discrete data, since the differences between multiple data are large, the use of lossy compression may lose a lot of data accuracy, so it can be stored according to its original value. Based on this, in some embodiments of this specification, the sample space data corresponding to the aforementioned vision examination data can be clustered using the K-means clustering method to obtain multiple result clusters, and then the compression method corresponding to the vision examination data in each result cluster is determined through subsequent steps.

[0072] Specifically, in some embodiments, the ISODATA algorithm can be used to adaptively cluster the vision examination data of all employees (i.e., the aforementioned sample space data) to obtain multiple result clusters. It should be noted that the ISODATA algorithm is an improved K-means clustering method, which can achieve better clustering effects by dynamically adjusting the number and position of clusters during the clustering process. The ISODATA algorithm is particularly suitable for clustering analysis when the number of clusters is unknown.

[0073] After the above clustering process, multiple result clusters can be obtained, wherein each result cluster represents a portion of vision examination data with similar characteristics, thereby dividing the overall sample data in the form of result clusters for subsequent calculations.

[0074] S140, calculating the data aggregation degree corresponding to each result cluster according to the data distribution characteristics of the result cluster, and merging adjacent result clusters at least according to the similarity of the data aggregation degrees to obtain a preferred cluster.

[0075] In some embodiments, the compression method corresponding to the vision examination data in each result cluster can be determined according to the data aggregation corresponding to each result cluster. In the embodiments of this specification, the data aggregation can represent the compactness of the data in a cluster. If the data aggregation of a cluster is larger, the data in the cluster is more compact, and it is more likely to use a lossy compression method for this cluster, while a cluster with a small data aggregation indicates that the data in it is relatively discrete, and it is more likely to use a lossless compression method.

[0076] Specifically, in some embodiments, the data aggregation degree corresponding to each result cluster can be calculated as follows:

[0077] First, the above result clusters can be processed using PCA (Principal Component Analysis) decomposition to obtain multiple principal component vectors and store them in sequence A; then, the morphological characteristics of each result cluster can be determined based on the modulus of the largest principal component vector in sequence A and the average modulus of all principal component vectors; further, the average density of data in each result cluster can be calculated based on the amount of data in the result cluster and the convex hull volume of the result cluster; finally, based on the above morphological characteristics, average data density, average distance between the data in the result cluster and the cluster center position, and average distance between the result cluster and adjacent clusters, the data aggregation degree corresponding to each result cluster can be determined.

[0078] Specifically, in some embodiments, sequence A can be expressed as (s is the number of principal component vectors obtained by decomposing this cluster). In some embodiments, the difference between the modulus of the largest principal component vector in sequence A and the average modulus of all principal component vectors can be used as the morphological feature of the corresponding result cluster (in terms of the morphology of the result cluster, the three-dimensional morphology of the clustered cluster is generally spherical, while the long strip cluster with an obvious morphological extension trend has a lower degree of aggregation); then, the morphological feature, the average density of the data, and the average distance between the data in the result cluster and the cluster center are multiplied to obtain a first product; finally, the ratio of the average distance between the result cluster and the adjacent clusters to the aforementioned first product is used as an index, and an exponential function is used for mapping to obtain the data aggregation degree corresponding to the result cluster.

[0079] In some embodiments, the aforementioned data aggregation degree can be calculated based on the following formula:

[0080]

[0081] Where Z is the data aggregation degree corresponding to the result cluster; exp{*} represents an exponential function with the natural constant e as the base, where * is the exponent; N is the amount of data in the result cluster; o represents the cluster center position in the result cluster, L(i,o) is the distance between the i-th data point in the cluster and the cluster center position o, and L(o,o) is the distance between the i-th data point in the cluster and the cluster center position o. j ) is the cluster center position o of the result cluster and the cluster center position o of the jth adjacent cluster j The distance; V is the convex hull volume of the result cluster; m is the number of adjacent clusters around the result cluster; is the modulus of the largest principal component vector in sequence A, is the average value of the modulus of all principal component vectors in sequence A. In this specification, the aforementioned distance may refer to the Euclidean distance.

[0082] Understandably, It represents the average distance between the data within the cluster and the center of the cluster. If the data within the result cluster is more clustered, the corresponding average distance is smaller, so the data clustering degree of the result cluster is greater. Represents the average distance between this result cluster and the adjacent clusters. If the data in this result cluster is more concentrated, the data is closer to the cluster center of this cluster and farther away from the cluster center of other clusters. The larger the value is, the more concentrated the data distribution in the result cluster is, and the greater the corresponding data aggregation degree is. Represents the average density of the data in this result cluster. If The smaller the value of, the greater the data density within the cluster, and therefore the greater the data aggregation. Represents the morphological characteristics of this result cluster. Since the clustered cluster has no obvious extension form, the modulus lengths of all principal component vectors after PCA decomposition are not much different. However, the non-clustered cluster has an obvious extension form, and the modulus lengths of all principal component vectors are quite different, and the modulus length of the largest principal component vector is It can indicate the degree of this extended form. The smaller the value is, the less obvious the extension form of this result cluster is, and the greater the data aggregation is.

[0083] As shown above, in some embodiments, the corresponding data aggregation degree can be calculated for all result clusters, and mapped using an exponential function (eg, an exponential function with the natural constant e as the base), thereby widening the difference.

[0084] Through the above steps, the corresponding data aggregation degree can be calculated for each result cluster. The larger the data aggregation degree Z is, the denser the data distribution of this result cluster is, and vice versa.

[0085] In some embodiments, considering that there may be some result clusters that are close to each other and contain less data and the corresponding data concentrations are different but very similar, in practical applications, it is hoped that these result clusters will be merged into a large cluster for compression, thereby reducing compression overhead and improving compression efficiency. Therefore, in some embodiments, adjacent result clusters can be merged at least according to the similarity of the aforementioned data concentrations to obtain a preferred cluster. In this specification, the preferred cluster can be understood as the data classification result obtained after clustering optimization of the result clusters obtained by the aforementioned clustering process.

[0086] Specifically, the process of merging adjacent result clusters to obtain the aforementioned preferred cluster can be referred to as follows:

[0087] For any target result cluster a, the similarity between the two can be first determined based on the difference between the data aggregation degree corresponding to the target result cluster a and the data aggregation degree corresponding to the adjacent result cluster b; then, the relative data volume of the adjacent result cluster b can be determined based on the data volume contained in the adjacent result cluster b and the maximum and minimum values ​​of the data volume contained in all result clusters; further, based on the Euclidean distance between the cluster center positions of the target result cluster a and the adjacent result cluster b, the aforementioned similarity and the aforementioned relative data volume, the possibility of merging the target result cluster a and the adjacent result cluster b is calculated; finally, according to the first preset threshold and the merging possibility calculated by the aforementioned process, the target result cluster a is merged with one or more adjacent result clusters b to obtain a preferred cluster.

[0088] In some embodiments, the Euclidean distance between the cluster center positions of the aforementioned target result cluster a and the adjacent result cluster b, the aforementioned similarity, and the relative data size can be multiplied to obtain a second product; then, the second product is normalized to obtain the possibility of merging the target result cluster a and the adjacent result cluster b.

[0089] In some embodiments, the calculation process of the possibility of merging the target result cluster a with the adjacent result cluster b can be expressed as follows:

[0090]

[0091] Among them, H(a,b) is the possibility of merging the target result cluster a with the adjacent result cluster b; N b is the amount of data contained in the adjacent result cluster b, max(N) is the maximum amount of data contained in all result clusters, and min(N) is the minimum amount of data contained in all result clusters; Z a With Z b are the data aggregation of the target result cluster a and the adjacent result cluster b respectively; L(o a ,o b ) is the Euclidean distance between the cluster center positions of the target result cluster a and the adjacent result cluster b. Represents the relative data size of the adjacent result cluster b. Since small clusters are expected to be merged in practical applications, the smaller the data size of result cluster b, the greater the possibility of merging it into result cluster i. a -Z b | represents the similarity of the data aggregation between the target result cluster a and the adjacent result cluster b. If the Z values ​​of the two result clusters are closer, the distribution characteristics of the data in the two result clusters are more similar, that is, |Z a -Z b The smaller the value of |, the greater the possibility of merging. Finally, the distance L(o a ,o b ), the closer the cluster centers of the two result clusters are, the greater the possibility of merging. In some embodiments, it is possible to a -Z b The value of | is used to calculate the possibility of merging the target result cluster a with the adjacent result cluster b.

[0092] As shown above, in some embodiments, the second product can be multiplied by a sigmoid function. or |Z a -Z b |Normalized to the range of [0,1], we can get the merging possibility of the target result cluster a and all adjacent result clusters.

[0093] In some embodiments, two result clusters whose merging probability is less than or equal to a first preset threshold value may be merged to obtain the aforementioned preferred cluster. As an example only, in some embodiments of this specification, the first preset threshold value may be 0.25.

[0094] S150: Determine a target compression method corresponding to each preferred cluster based on the data concentration and compression loss of the preferred cluster.

[0095] For the preferred cluster, if the corresponding data aggregation Z value is smaller, it means that the data in this cluster is more aggregated, then the data loss during lossy compression is smaller, and it is more suitable for lossy compression storage. Conversely, it is more suitable for lossless compression storage. Among them, lossy compression can be understood as taking the mean of all data points in the preferred cluster, and then using this mean to replace each data point in the cluster, so as to achieve the purpose of lossy compression and saving storage space by using the principle of run-length encoding. Lossless compression can be understood as storing according to its original value.

[0096] Based on this, for each target preferred cluster, its corresponding data aggregation can be calculated first; then, the relative data aggregation of the target preferred cluster can be determined based on the maximum value of the data aggregation corresponding to the target preferred cluster and the data aggregation corresponding to all preferred clusters; further, the compression coefficient corresponding to the target preferred cluster can be calculated based on the aforementioned relative data aggregation and the cumulative sum of the differences between each data in the target preferred cluster and the average value of the data in the cluster; finally, the target compression method corresponding to the target preferred cluster can be determined based on the compression coefficient.

[0097] Specifically, in some embodiments of the present specification, the average data aggregation corresponding to all result clusters merged by the target preferred cluster can be calculated, and the average data aggregation can be used as the data aggregation corresponding to the target preferred cluster. In some embodiments, the aforementioned relative data aggregation can be multiplied by the cumulative sum of the difference between each data in the target preferred cluster and the average value of the data in the cluster to obtain a third product, and the third product can be used as the compression coefficient corresponding to the target preferred cluster.

[0098] In some embodiments, the calculation process of the compression coefficient corresponding to the target preferred cluster t can be expressed as follows:

[0099]

[0100] Among them, R t is the compression coefficient corresponding to the target optimal cluster t; is the average value of all data points in the target optimal cluster t, (x k ,y k ,z k ) is the kth data point in the target optimal cluster t; N t is the amount of data contained in the target optimal cluster t, Z t is the data aggregation degree corresponding to the target preferred cluster t, and max(Z) is the maximum value of the data aggregation degrees corresponding to all preferred clusters. It is the cumulative sum of the differences between each data in the target preferred cluster t and the average value of the data in the cluster, which represents the data loss when lossy compression is performed on this cluster. The greater the data loss, the greater the compression factor, and the more likely it is that lossless compression is required. Represents the relative data concentration of the target preferred cluster t. If Z t The smaller the cluster, the more discrete it is. The larger the value, the greater the compression factor, and the more likely it is that lossless compression will be required.

[0101] Furthermore, the compression coefficient calculated in the above process may be normalized to obtain a normalized result, and then the target compression method corresponding to the target preferred cluster may be determined according to the normalized result.

[0102] Specifically, in some embodiments, when the normalized result corresponding to the target preferred cluster is less than or equal to the second preset threshold, the target compression mode corresponding to the target preferred cluster can be determined as lossy compression; when the normalized result corresponding to the target preferred cluster is greater than the second preset threshold, the target compression mode corresponding to the target preferred cluster can be determined as lossless compression. As an example only, in some embodiments of this specification, the second preset threshold can be 0.3.

[0103] S160, compressing and storing the vision examination data in the corresponding preferred cluster according to the target compression method.

[0104] After the above steps, the target compression method corresponding to each preferred cluster can be obtained. Furthermore, the visual acuity examination data in the corresponding preferred cluster can be compressed and stored according to the target compression method, that is, different compression methods can be used for efficient storage according to the data aggregation degree and data loss of the preferred cluster, thereby saving storage space while ensuring data accuracy and improving the management efficiency of the ERP management system.

[0105] It should be noted that when lossy compression is performed on the employee's vision examination data, the employee's name needs to be accurately matched with his / her vision examination data, so as to ensure that the employee's vision information can be accurately obtained during the subsequent data reading process.

[0106] In summary, the beneficial effects that may be brought about by the embodiments of the present specification include but are not limited to: (1) In the data processing method of the physical examination ERP management system provided in some embodiments of the present specification, by clustering and merging the sample space data corresponding to the employee vision examination data, one or more preferred clusters are obtained, and then different compression storage methods are determined by calculating the aggregation degree and compression loss of the preferred clusters, which can improve the redundant space caused by direct storage, thereby saving storage space and improving the storage efficiency of the ERP management system; (2) In the data processing method of the physical examination ERP management system provided in some embodiments of the present specification, by determining the relative data aggregation degree according to the data aggregation degree corresponding to the target preferred cluster and the maximum value of the data aggregation degrees corresponding to all preferred clusters, and then according to the relative data aggregation degree and the difference between each data in the target preferred cluster and the data in the cluster, the relative data aggregation degree is calculated. The cumulative sum of the differences between the average values ​​is used to calculate the compression coefficient corresponding to the target preferred cluster, which can not only take into account the clustering degree and compression loss of the preferred cluster at the same time, but also amplify the compression loss corresponding to the target preferred cluster when lossy compression is used by calculating the cumulative sum, so that its loss characteristics are more obvious, thereby improving the accuracy of the target compression method determined based on this; (3) In the data processing method of the physical examination ERP management system provided in some embodiments of the present specification, by normalizing the compression coefficient corresponding to the target preferred cluster, and then when the corresponding normalized result is less than or equal to the second preset threshold, the corresponding target compression method is determined as lossy compression, and when the corresponding normalized result is greater than the second preset threshold, the corresponding target compression method is determined as lossless compression, which can save storage space while ensuring data accuracy and reliability.

[0107] It should be noted that different embodiments may produce different beneficial effects. In different embodiments, the beneficial effects that may be produced may be any one or a combination of the above, or may be any other possible beneficial effects.

[0108] The basic concepts have been described above. Obviously, for those skilled in the art, the above detailed disclosure is only for example and does not constitute a limitation of this specification. Although not explicitly stated here, those skilled in the art may make various modifications, improvements and corrections to this specification. Such modifications, improvements and corrections are suggested in this specification, so such modifications, improvements and corrections still belong to the scope of the exemplary embodiments of this specification.

Claims

1. A data processing method for a physical examination ERP management system, characterized in that: include: Obtain employees' vision examination data and establish sample space; Mapping the vision examination data to the sample space to obtain a plurality of sample space data; Clustering the sample space data using the ISODATA algorithm to obtain multiple result clusters; Calculating the data aggregation degree corresponding to each result cluster according to the data distribution characteristics of the result cluster, and merging adjacent result clusters at least according to the similarity of the data aggregation degrees to obtain a preferred cluster; Determining a target compression method corresponding to each preferred cluster based on the data concentration and compression loss of the preferred cluster, wherein the target compression method includes lossy compression and lossless compression; Compressing and storing the vision examination data in the corresponding preferred cluster according to the target compression method; The calculating the data aggregation degree corresponding to each result cluster according to the data distribution characteristics of the result cluster includes: Process the result clusters using PCA decomposition to obtain multiple principal component vectors, and store them in sequence A; Determining the morphological characteristics of the result cluster according to the modulus of the largest principal component vector in the sequence A and the average of the modulus of all principal component vectors; Calculating the average density of the data in the result cluster according to the amount of data in the result cluster and the convex hull volume of the result cluster; Determine the data aggregation degree corresponding to the result cluster based on the morphological features, the average density of the data, the average distance between the data in the result cluster and the cluster center position, and the average distance between the result cluster and adjacent clusters; The step of determining the data aggregation degree corresponding to the result cluster based on the morphological features, the average density of the data, the average distance between the data in the result cluster and the cluster center position, and the average distance between the result cluster and adjacent clusters includes: The difference between the modulus of the largest principal component vector in the sequence A and the average value of the modulus of all principal component vectors is used as the morphological feature of the result cluster; Multiplying the morphological feature, the average data density, and the average distance between the data in the result cluster and the cluster center position to obtain a first product; The ratio of the average distance between the result cluster and the adjacent clusters to the first product is used as an index, and an exponential function is used for mapping to obtain the data aggregation degree corresponding to the result cluster; The step of merging adjacent result clusters at least according to the similarity of the data aggregation degree to obtain a preferred cluster includes: For any target result cluster a; Determine the similarity between the target result cluster a and the adjacent result cluster b according to the difference between the data aggregation degree corresponding to the target result cluster a and the data aggregation degree corresponding to the adjacent result cluster b; Determine the relative data size of the adjacent result cluster b according to the data size of the adjacent result cluster b and the maximum and minimum values ​​of the data sizes of all result clusters; Calculate the possibility of merging the target result cluster a and the adjacent result cluster b based on the Euclidean distance between the cluster center positions of the target result cluster a and the adjacent result cluster b, the similarity, and the relative data size; Merging the target result cluster a with one or more adjacent result clusters b according to a first preset threshold and the merging possibility to obtain the preferred cluster; The calculating the possibility of merging the target result cluster a and the adjacent result cluster b based on the Euclidean distance between the cluster center positions of the target result cluster a and the adjacent result cluster b, the similarity, and the relative data size includes: Multiplying the Euclidean distance between the cluster center positions of the target result cluster a and the adjacent result cluster b, the similarity, and the relative data volume to obtain a second product; Normalizing the second product to obtain a possibility of merging the target result cluster a with the adjacent result cluster b; The determining of a target compression method corresponding to each preferred cluster based on the data concentration and compression loss of the preferred cluster includes: For each target preferred cluster; Calculating the data aggregation degree corresponding to the target preferred cluster; Determining the relative data aggregation degree of the target preferred cluster according to the data aggregation degree corresponding to the target preferred cluster and the maximum value of the data aggregation degrees corresponding to all preferred clusters; Calculate the compression coefficient corresponding to the target preferred cluster according to the relative data aggregation and the cumulative sum of the difference between each data in the target preferred cluster and the average value of the data in the cluster; Determining a target compression mode corresponding to the target preferred cluster according to the compression coefficient; The calculating the data aggregation degree corresponding to the preferred cluster includes: Calculate the average data aggregation degree corresponding to all result clusters merged by the target preferred cluster; Using the average data aggregation degree as the data aggregation degree corresponding to the target preferred cluster; The calculation formula for data aggregation is: Where Z is the data aggregation degree corresponding to the result cluster; exp{*} represents an exponential function with the natural constant e as the base, where * is the exponent; N is the amount of data in the result cluster; o represents the cluster center position in the result cluster, L(i,o) is the distance between the i-th data point in the cluster and the cluster center position o, and L(o,o) is the distance between the i-th data point in the cluster and the cluster center position o. j ) is the cluster center position o of the result cluster and the cluster center position o of the jth adjacent cluster j The distance; V is the convex hull volume of the result cluster; m is the number of adjacent clusters around the result cluster; is the modulus of the largest principal component vector in sequence A, is the average value of the modulus of all principal component vectors in sequence A; The calculating the compression coefficient corresponding to the target preferred cluster according to the relative data concentration and the cumulative sum of the difference between each data in the target preferred cluster and the average value of the data in the cluster includes: Multiplying the relative data concentration by the cumulative sum of the differences between each data in the target preferred cluster and the average value of the data in the cluster to obtain a third product; The third product is used as the compression coefficient corresponding to the target preferred cluster.

2. The data processing method of the physical examination ERP management system according to claim 1, characterized in that: The determining, according to the compression coefficient, a target compression mode corresponding to the target preferred cluster includes: Normalizing the compression coefficient to obtain a normalized result; When the normalized result corresponding to the target preferred cluster is less than or equal to a second preset threshold, determining the target compression mode corresponding to the target preferred cluster as lossy compression; When the normalized result corresponding to the target preferred cluster is greater than the second preset threshold, the target compression mode corresponding to the target preferred cluster is determined to be lossless compression.

3. The data processing method of a physical examination ERP management system according to claim 2, characterized in that: The first preset threshold is 0.25, and the second preset threshold is 0.3.

Images