Method, medium and device for obtaining delivery sites based on keywords and spatial clustering

Abstract

The present application discloses a method for obtaining a delivery station based on keywords and spatial clustering, and relates to the field of express delivery. By using historical address data to screen out effective frequent items to construct a keyword table, the address keywords of the address to be assigned and the longitude and latitude corresponding to the address keywords can be automatically extracted according to the keyword table, reducing the manual maintenance cost; one address keyword corresponds to only one station, which improves the accuracy of express station allocation; extracting keywords from the address, retaining the key information of the address, and increasing the repetition rate of similar addresses, so that the longitude and latitude can be corrected by big data clustering, thereby further improving the accuracy of express station allocation and reducing transportation costs.

Description

Technical Field

[0001] The present application relates to the field of express delivery, and in particular to a method, medium and device for acquiring delivery sites based on keywords and spatial clustering. Background Art

[0002] After the express is connected to the network, the terminal delivery site needs to be assigned according to the relevant information such as the delivery address filled in by the user. In some areas, postal codes also play an important role in distinguishing geographical areas, and sites can be assigned based on postal codes. However, in some areas, there are many postal codes and there are mixed postal codes of different sizes and old and new postal codes. One postal code may correspond to multiple sites, so the workload of maintaining the postal code site correspondence table is very large, and it is impossible to accurately determine the corresponding site.

[0003] Some logistics companies also obtain their locations through external map vendors or through the historical delivery data of small parcel workers. However, the longitude and latitude of the addresses obtained by external map vendors are inaccurate, and the delivery location data obtained by small parcel workers is inconsistent with the addresses, resulting in point drift, which in turn leads to inaccurate sites. We have also tried to integrate external map vendors and delivery data of small parcel workers to correct the address location, but this method can only be used for addresses that have appeared and are completely consistent, the amount of data is small, and the effect is not ideal. Summary of the invention

[0004] In order to solve at least one of the technical problems mentioned in the background technology, the purpose of the present application is to provide a method, medium and device for obtaining delivery sites based on keywords and spatial clustering, which can improve the allocation accuracy of delivery sites and reduce express transportation costs.

[0005] To achieve the above objectives, this application provides the following technical solutions:

[0006] In a first aspect, an embodiment of the present application provides a method for acquiring a delivery site based on keywords and spatial clustering, comprising the following steps:

[0007] Pre-building a keyword table includes the following steps:

[0008] S1, based on multiple historical address data, construct an address data table, which contains administrative divisions, postal codes, detailed addresses, and address latitude and longitude;

[0009] S2, segmenting the detailed address to obtain an address data set; counting the frequencies of elements in the address data set, sorting the frequencies from large to small, selecting elements whose ranking digits are before a first preset threshold as key elements, and scoring them based on the frequencies and meanings, and constructing a key element table using the key elements and the corresponding frequencies and scores;

[0010] S3, combining the elements in the address data set to obtain multiple subsets; taking the subsets whose first element is the key element as frequent items to form a frequent item set;

[0011] S4, matching the subset corresponding to the address data set of each piece of historical address data with the frequent items in the frequent item set, and for the frequent items with inclusion relationship, removing the included frequent items to obtain the valid frequent items corresponding to the historical address data;

[0012] S5, for each historical address data, count the frequency of the effective frequent items under the corresponding administrative division and postal code, and calculate the sum of the frequencies of the effective frequent items; use the effective frequent items and the corresponding sum of frequencies to combine with the address data table to construct an original keyword table; wherein the set of effective frequent items is used as the address keyword to replace the detailed address;

[0013] S6, for the original keyword table, grouping the same administrative divisions, postal codes, and address keywords to obtain multiple groups, removing the groups that do not meet the preset requirements, calculating the center longitude and latitude of the remaining groups as the new longitude and latitude corresponding to the address keyword; and calculating the sum of the scores of the key elements in the address keyword as the score of the address keyword; using the administrative divisions, postal codes, address keywords, and the corresponding new longitude and latitude and scores to construct a keyword table;

[0014] Obtaining the address to be allocated, wherein the address to be allocated includes an administrative division, a postal code, and a detailed address, and obtaining multiple subsets corresponding to the address to be allocated by using steps S2 to S3;

[0015] According to the administrative division and postal code, searching for all corresponding address keywords from the pre-constructed keyword table;

[0016] Filter out new longitudes and latitudes corresponding to address keywords with the highest scores that can be covered by multiple subsets corresponding to the address to be assigned from all determined address keywords as the longitudes and latitudes of the address to be assigned;

[0017] Use the longitude and latitude of the address to be assigned to mark a point on the site map to obtain the corresponding delivery site.

[0018] Furthermore, the step of segmenting the detailed address includes:

[0019] Convert administrative divisions and detailed addresses into lowercase letters;

[0020] According to the standard administrative division mapping table, the administrative divisions are mapped to standard administrative divisions;

[0021] Clean the detailed address and remove meaningless symbols;

[0022] The detailed address is split according to spaces and commas to obtain an address dataset consisting of single words.

[0023] Furthermore, before counting the frequencies of elements in the address data set, numbers, symbols, and single letters in the address data set are removed.

[0024] Furthermore, the step of combining the elements in the address data set includes:

[0025] The elements in the address data set are combined in sequence using the incremental construction method to obtain multiple subsets;

[0026] The number of elements in the subset is greater than or equal to 2 and less than or equal to 6.

[0027] Further, after taking the subset whose first element is the key element as the frequent item, the method further includes:

[0028] Counting the frequencies of subsets whose first elements are key elements, and removing subsets whose frequencies are less than or equal to a second preset threshold;

[0029] The remaining subsets whose first elements are key elements are taken as frequent items to form a frequent item set.

[0030] Furthermore, the step of removing the groups that do not meet the preset requirements includes:

[0031] For each group, cluster the nodes in the group using a clustering algorithm to obtain multiple clusters, each cluster containing a cluster center; wherein the node refers to the data corresponding to each address keyword;

[0032] Calculate the longitude and latitude of each cluster center according to the longitude and latitude corresponding to the address keyword, and calculate the distance between the cluster centers; if the farthest distance between the cluster centers is greater than a third preset threshold, remove the data of the group;

[0033] Calculating the proportion of outliers, and if the proportion of outliers is greater than a fourth preset threshold, removing the data of the group;

[0034] According to the longitude and latitude of the nodes in the group, points are placed on the site map. If the nodes with latitude and longitude less than a fifth preset threshold belong to and only belong to the same site, the data of the group is removed.

[0035] Furthermore, after removing the groups that do not meet the preset requirements, the method further includes:

[0036] In the remaining groups, the data judged as outliers are removed.

[0037] In a second aspect, an embodiment of the present application provides a computer storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned method for acquiring delivery sites based on keywords and spatial clustering.

[0038] In a third aspect, an embodiment of the present application provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements a method for acquiring delivery sites based on keywords and spatial clustering when executing the computer program.

[0039] Compared with the prior art, the beneficial effects of the present application are: by using historical address data to screen out effective frequent items to construct a keyword table, the address keywords of the address to be assigned and the longitude and latitude corresponding to the address keywords can be automatically extracted according to the keyword table, reducing manual maintenance costs; one address keyword only corresponds to one site, which improves the accuracy of express site allocation; by extracting keywords from the address, while retaining the key information of the address, the repetition rate of similar addresses is increased, so that the longitude and latitude can be corrected through big data clustering, thereby further improving the accuracy of express site allocation and reducing transportation costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A flowchart of a method for constructing a keyword table provided in Example 1 of the present application;

[0041] Figure 2 A flowchart of a method for acquiring delivery sites based on keywords and spatial clustering provided in Example 1 of the present application. DETAILED DESCRIPTION

[0042] The technical solutions in the embodiments of the present application are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.

[0043] In order to solve the above problems existing in the prior art, the embodiments of the present application disclose a method, medium and device for obtaining a delivery site based on keywords and spatial clustering. It should be noted that the execution subject of the embodiments of the present application can be a computer or any electronic device with data processing capabilities. For ease of description, the following detailed description is given with the electronic device as the execution subject.

[0044] Embodiment 1:

[0045] like Figure 1 As shown, Figure 1A flowchart of a method for constructing a keyword table provided in Example 1 of the present application. The method includes:

[0046] S1, based on multiple historical address data, construct an address data table, which contains administrative divisions, postal codes, detailed addresses, and address latitude and longitude.

[0047] Obtain address-related data from January 1, 2023 to December 31, 2023 from the geographic database, extract the postal code, first- and second-level administrative divisions in the address, split the address, and construct an address data table, as shown in Table 1 below.

[0048] Table 1: Address data table

[0049]

[0050] Convert the character strings in the first-level administrative divisions, second-level administrative divisions, and detailed addresses in Table 1 into lowercase.

[0051] Since these strings are filled in by users, there may be some spelling errors or irregular formats. The standard administrative division mapping table is obtained from the system to map the first-level administrative divisions and the second-level administrative divisions to standard administrative divisions.

[0052] Clean the detailed address, remove meaningless symbols and reduce data noise.

[0053] S2, segment the detailed address to obtain an address data set; count the frequencies of elements in the address data set, sort the frequencies from large to small, select elements whose ranking digits are before the first preset threshold as key elements, and score them based on the frequency and meaning, and use the key elements and the corresponding frequencies and scores to construct a key element table.

[0054] The detailed address is split according to spaces and commas, and the detailed address is converted into an address data set consisting of single words, as shown in Table 2 below.

[0055] Table 2: Address dataset table

[0056]

[0057] Extract individual elements from the address data set, remove numbers, symbols, and single letters, count the frequency of the elements, sort them from high to low by frequency, extract the elements with the top 300 frequencies, filter out elements related to the address, score them based on the frequency and the meaning of the element words, and construct a key element table, as shown in Table 3 below.

[0058] Table 3: Key elements table

[0059]

[0060] In Table 3, the smaller the frequency of an element, the higher its accuracy is considered, so the score will be higher. In addition, from the meaning point of view, the scope of the road is relatively wide, and the garden is equivalent to the community in China, so the garden is more accurate and the score will be higher.

[0061] S3, combining the elements in the address data set to obtain multiple subsets; taking the subsets whose first elements are key elements as frequent items to form a frequent item set.

[0062] The address data set is split and combined into multiple subsets by using the incremental construction method in element order. The number of elements in a subset is at least 2 elements and at most 6 elements. The number of repeated occurrences of a subset with more than 6 elements is too small, and the number of repeated occurrences of a subset with 1 element is too large.

[0063] For example, the address dataset is: {1-29, tingkat, 1, blok, rps, jalan, bukit, permai, 2, taman, bukit, permai, ampang}, total 13 elements.

[0064] The subsets obtained by combination are: {1 - 29, tingkat}; {1 - 29, tingkat, 1}; {1 - 29, tingkat, 1, blok}; {1 - 29, tingkat, 1, blok, rps}; {1 - 29, tingkat, 1, blok, rps, jalan}; {tingkat, 1}; {tingkat, 1, blok}; {tingkat, 1, blok, rps}; {tingkat, 1, blok, rps, jalan}; {tingkat, 1, blok, rps, jalan, bukit}; {1, blok}; {1, blok, rps}; {1, blok, rps, jalan}; {1, blok, rps, jalan, bukit}; {1, blok, rps, jalan, bukit, permai}; {blok, rps}; {blok, rps, jalan}; {blok, rps, jalan, bukit}; {blok, rps, jalan, bukit, permai}; {blok, rps, jalan, bukit, permai, 2}; {rps, jalan}; {rps, jalan, bukit}; {rps, jalan, bukit, permai}; {rps, jalan, bukit, permai, 2}; {rps, jalan, bukit, permai, 2, taman}; {jalan, bukit}; {jalan, bukit, permai}; {jalan, bukit, permai, 2}; {jalan, bukit, permai, 2, taman}; {jalan, bukit, permai, 2, taman, bukit}; {bukit, permai}; {bukit, permai, 2}; {bukit, permai, 2, taman}; {bukit, permai, 2, taman, bukit}; {bukit, permai, 2, taman, bukit, permai}; {permai, 2}; {permai, 2, taman}; {permai, 2, taman, bukit}; {permai, 2, taman, bukit, permai}; {permai, 2, taman, bukit, permai, ampang}; {2, taman}; {2, taman, bukit}; {2, taman, bukit, permai}; {2, taman, bukit, permai, ampang};{taman, bukit}; {taman, bukit, permai}; {taman, bukit, permai, ampang}; {bukit, permai}; {bukit, permai, ampang}; {permai, ampang}; a total of 50 subsets. ;

[0065] The subsets obtained above are filtered to retain the subsets whose first elements are in the key element table.

[0066] In some implementations, in order to improve the accuracy of selecting frequent items, the above subsets can be further screened to retain only subsets with frequencies greater than 10. The frequency threshold of 10 is only used as an example. The frequency threshold can be set according to actual needs, but it should not be too high. Setting the frequency threshold too high will result in the loss of more frequent items, which is not limited in this application.

[0067] The remaining subsets after filtering are regarded as frequent items to form a frequent item set. Taking the above case as an example, the frequent items are: {jalan, bukit, permai, 2}, {taman, bukit, permai, ampang}, {jalan, bukit, permai}, {taman, bukit, permai}.

[0068] S4, matching the subset corresponding to the address data set of each piece of historical address data with the frequent items in the frequent item set, and for frequent items with inclusion relationship, eliminating the included frequent items to obtain the valid frequent items corresponding to the historical address data.

[0069] For each detailed address with administrative divisions and postal codes, its subset is matched with the frequent items in the frequent item set to obtain all frequent items corresponding to the detailed address; for two frequent items with inclusion relationship, the largest item is retained and the included frequent items are eliminated, thereby obtaining the valid frequent items corresponding to the detailed address.

[0070] S5, for each historical address data, count the frequency of valid frequent items under the corresponding administrative division and postal code, and calculate the sum of the frequencies of the valid frequent items; use the valid frequent items and the corresponding sum of the frequencies to combine with the address data table to construct the original keyword table; among them, the set of valid frequent items is used as the address keyword to replace the detailed address.

[0071] On the basis of Table 1, the detailed address is replaced by the set of calculated effective frequent items, a new frequency column is added, the value of which is the sum of the frequencies of effective frequent items, and the data with a frequency greater than 5 is retained to obtain the original keyword table, as shown in Table 4 below.

[0072] Table 4: Original keyword list

[0073]

[0074] S6, for the original keyword table, group the administrative divisions, postal codes, and address keywords to obtain multiple groups, remove the groups that do not meet the preset requirements, calculate the central longitude and latitude of the remaining groups, and use them as the new longitude and latitude corresponding to the address keyword; and calculate the sum of the scores of the key elements in the address keyword as the score of the address keyword; use the administrative divisions, postal codes, address keywords and the corresponding new longitude and latitude and scores to construct a keyword table.

[0075] Each data in the original keyword table is grouped, and the data with the same administrative division, postal code and address keywords are grouped together, thus obtaining multiple groups.

[0076] For each group, each piece of data in the group (that is, each row of data in the table) is regarded as a node, and clustering is performed using the DBSCAN clustering algorithm. The minimum number of clustering points is set to 3, and the longitude and latitude distance is set to 2km for clustering. Multiple clusters are obtained, and each cluster has a cluster center.

[0077] Calculate the percentage of outliers. Outliers are points that are considered noise during the DBSCAN clustering process, that is, points that do not belong to any cluster. If more than 20% of the points in a group are considered outliers, it may mean that the address keyword range in this group is too large and not precise enough.

[0078] Calculate the longitude and latitude of each cluster center, and calculate the maximum distance between cluster centers based on the longitude and latitude of the cluster center. If the maximum distance between cluster centers is greater than 3km, it may be because the addresses represented by these cluster centers are too geographically dispersed to reasonably assume that they belong to the same address or location. This may indicate that the scope of the keyword frame is too broad, for example, a very long road is extracted, and this data can be judged as insufficiently accurate.

[0079] Get the site map from the geographic database, and mark the nodes in the group on the site map. If 95% or more of the points belong to and only belong to one site, the site filter column of the group is marked as 1, otherwise it is marked as 0. If the site filter column is marked as 0, it means that although these points belong to a group of administrative divisions, postal codes, and detailed address keywords, when marking on the site map, these points do not belong to and only belong to 95% or more of the sites. This usually indicates that the address keyword spans multiple sites and the keyword range is not precise enough.

[0080] Therefore, it is necessary to remove the groups in the original keyword table whose cluster centers have a maximum distance greater than 3 km or whose site filter column is marked as 0 or whose outlier ratio is greater than 20%. This can improve the accuracy of the keywords.

[0081] In the remaining groups, the data judged as outliers are eliminated, and the central longitude and latitude of each group of data are calculated. It can be calculated based on the average longitude and latitude of all nodes or the average longitude and latitude of the cluster center. The calculated central longitude and latitude are used as the new longitude and latitude of the group of data, and the sum of the scores of the key elements contained in the address keyword is calculated as the score of the group of data, thereby obtaining a keyword table as shown in Table 5 below.

[0082] Table 5: Keyword table

[0083]

[0084] like Figure 2 As shown, Figure 2 A flowchart of a method for obtaining a delivery site based on keywords and spatial clustering provided in Example 1 of the present application. The method includes:

[0085] The address to be allocated is obtained, wherein the address to be allocated includes an administrative division, a postal code, and a detailed address, and multiple subsets corresponding to the address to be allocated are obtained by using steps S2 to S3 in the first embodiment.

[0086] According to the administrative division and postal code, all corresponding address keywords are searched from the keyword table constructed in the first embodiment.

[0087] In the keyword table, a pair of administrative divisions and postal codes corresponds to at least one address keyword. Since most of the noise has been removed during the construction of the keyword table, the keyword search scope can be greatly reduced.

[0088] Among all the determined address keywords, the new longitude and latitude corresponding to the address keyword with the highest score that can be covered by the multiple subsets corresponding to the address to be assigned are selected as the longitude and latitude of the address to be assigned.

[0089] Only address keywords that can be covered by the subset corresponding to the address to be assigned can match the address to be assigned. This step may screen out multiple address keywords. According to the score in the last column of the keyword table, the new longitude and latitude corresponding to the address keyword with the highest score can be selected as the longitude and latitude of the address to be assigned.

[0090] Use the longitude and latitude of the address to be assigned to mark a point on the site map to obtain the corresponding delivery site.

[0091] The delivery site acquisition method based on keywords and spatial clustering provided in the present application uses historical address data to screen out effective frequent items to construct a keyword table, and can automatically extract the address keywords of the address to be assigned and the longitude and latitude corresponding to the address keywords according to the keyword table, thereby reducing manual maintenance costs; one address keyword corresponds to only one site, which improves the accuracy of express site allocation; keywords are extracted from the address, while retaining the key information of the address, the repetition rate of similar addresses is increased, so that the longitude and latitude can be corrected through big data clustering, thereby further improving the accuracy of express site allocation; the longitude and latitude distance is used in the DBSCAN clustering algorithm, so that the parameters have actual physical meaning, the parameter adjustment is more convenient, and the interpretability is stronger.

[0092] Embodiment 2:

[0093] Corresponding to the embodiment of the method for acquiring delivery sites based on keywords and spatial clustering, the embodiment of the present application also provides a computer storage medium on which a computer program is stored, which implements the above-mentioned method for acquiring delivery sites based on keywords and spatial clustering when executed by a processor.

[0094] Embodiment three:

[0095] Corresponding to the embodiment of the method for acquiring delivery sites based on keywords and spatial clustering, the embodiment of the present application also provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the above-mentioned method for acquiring delivery sites based on keywords and spatial clustering when executing the computer program.

[0096] It is obvious to those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and that the present application can be implemented in other specific forms without departing from the spirit or essential features of the present application. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive, and the scope of the present application is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present application.

Claims

1. A method for obtaining delivery sites based on keywords and spatial clustering, characterized in that: The following steps are involved: Pre-building a keyword table includes the following steps: S1, based on multiple historical address data, construct an address data table, which contains administrative divisions, postal codes, detailed addresses, and address latitude and longitude; S2, segmenting the detailed address to obtain an address data set; counting the frequencies of elements in the address data set, sorting the frequencies from large to small, selecting elements whose ranking digits are before a first preset threshold as key elements, and scoring them based on the frequencies and meanings, and constructing a key element table using the key elements and the corresponding frequencies and scores; S3, combining the elements in the address data set to obtain multiple subsets; taking the subsets whose first element is the key element as frequent items to form a frequent item set; S4, matching the subset corresponding to the address data set of each piece of historical address data with the frequent items in the frequent item set, and for the frequent items with inclusion relationship, removing the included frequent items to obtain the valid frequent items corresponding to the historical address data; S5, for each historical address data, count the frequency of the effective frequent items under the corresponding administrative division and postal code, and calculate the sum of the frequencies of the effective frequent items; use the effective frequent items and the corresponding sum of frequencies to combine with the address data table to construct an original keyword table; wherein the set of effective frequent items is used as the address keyword to replace the detailed address; S6, for the original keyword table, grouping the same administrative divisions, postal codes, and address keywords to obtain multiple groups, removing the groups that do not meet the preset requirements, calculating the center longitude and latitude of the remaining groups as the new longitude and latitude corresponding to the address keyword; and calculating the sum of the scores of the key elements in the address keyword as the score of the address keyword; using the administrative divisions, postal codes, address keywords, and the corresponding new longitude and latitude and scores to construct a keyword table; Obtaining the address to be allocated, wherein the address to be allocated includes an administrative division, a postal code, and a detailed address, and obtaining multiple subsets corresponding to the address to be allocated by using steps S2 to S3; According to the administrative division and postal code, searching for all corresponding address keywords from the pre-constructed keyword table; Filter out new longitudes and latitudes corresponding to address keywords with the highest scores that can be covered by multiple subsets corresponding to the address to be assigned from all determined address keywords as the longitudes and latitudes of the address to be assigned; Use the longitude and latitude of the address to be assigned to mark a point on the site map to obtain the corresponding delivery site.

2. The method for obtaining delivery sites based on keywords and spatial clustering according to claim 1, characterized in that: The step of segmenting the detailed address includes: Convert administrative divisions and detailed addresses into lowercase letters; According to the standard administrative division mapping table, the administrative divisions are mapped to standard administrative divisions; Clean the detailed address and remove meaningless symbols; The detailed address is split according to spaces and commas to obtain an address dataset consisting of single words.

3. The method for obtaining delivery sites based on keywords and spatial clustering according to claim 1, characterized in that: Before counting the frequencies of elements in the address data set, numbers, symbols, and single letters in the address data set are removed.

4. The method for obtaining delivery sites based on keywords and spatial clustering according to claim 1, characterized in that: The step of combining the elements in the address data set includes: The elements in the address data set are combined in sequence using the incremental construction method to obtain multiple subsets; The number of elements in the subset is greater than or equal to 2 and less than or equal to 6.

5. The method for obtaining delivery sites based on keywords and spatial clustering according to claim 1, characterized in that: After taking the subset whose first element is the key element as the frequent item, the method further includes: Counting the frequencies of subsets whose first elements are key elements, and removing subsets whose frequencies are less than or equal to a second preset threshold; The remaining subsets whose first elements are key elements are taken as frequent items to form a frequent item set.

6. The method for obtaining delivery sites based on keywords and spatial clustering according to claim 1, characterized in that: The step of removing the groups that do not meet the preset requirements includes: For each group, cluster the nodes in the group using a clustering algorithm to obtain multiple clusters, each cluster containing a cluster center; wherein the node refers to the data corresponding to each address keyword; Calculate the longitude and latitude of each cluster center according to the longitude and latitude corresponding to the address keyword, and calculate the distance between the cluster centers; if the farthest distance between the cluster centers is greater than a third preset threshold, remove the data of the group; Calculating the proportion of outliers, and if the proportion of outliers is greater than a fourth preset threshold, removing the data of the group; According to the longitude and latitude of the nodes in the group, points are placed on the site map. If the nodes with latitude and longitude less than a fifth preset threshold belong to and only belong to the same site, the data of the group is removed.

7. The method for obtaining delivery sites based on keywords and spatial clustering according to claim 6, characterized in that: After removing the groups that do not meet the preset requirements, the method further includes: In the remaining groups, the data judged as outliers are removed.

8. A computer storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

Images