A classification method and a classification device for service data

2:

• The classification method according to claim 1, wherein the extracting comprises:
• extracting the service data, of which a certain service indicator data is an outlier according to a set threshold to form a first data set;
• or,
• extracting, according to a logical operation result of a plurality of service indicators of the service data, the service data, of which the logical operation result is an outlier to form a first data set;
• or,
• extracting, according to a data distribution of a certain service indicator, service data of which the certain service indicator data is an outlier, to form a first data set.

4:

• The classification method according to claim 3, wherein the merging comprises:
• determining an average feature value of service indicators involved in the extraction rule in the first data set and the second data set, and merging the classes in the first data set with the classes in the second data set having the closest category attributes to the classes in the first data set to obtain the classification result of the service data, wherein the average feature value is an average value or a median of the service indicators involved in the extraction rule of each class;
• or,
• merging the classes in the first data set with the classes in the second data set having the closest category attributes to the classes in the first data set to obtain the classification result of the service data, and making the difference in the number of the service data between each service data classification in the classification result conforms to a preset range.

6:

• The classification method according to claim 1, wherein further comprising, prior to clustering the service data in the second data set:
• selecting service indictor data that are outliers from the second data set, and assigning a service indicator demarcation value used to determine whether the service indicator data are outliers to the service indicator data that are outliers;
• or,
• calculating a mean value of all non-null data of a certain service indicator of the second data set, and assigning the mean value to null values of the service indicator in the second data set.

7:

• The classification method according to claim 1, wherein the clustering comprises:
• pre-clustering the service data in the second data set according to each predetermined number of clusters, respectively, and calculating a profile coefficient of a pre-clustering result corresponding to each predetermined number of clusters;
• ranking the corresponding profile coefficients in an ascending order of the predetermined number of clusters, and acquiring several peak values in the profile coefficients and determining a maximum value therefrom;
• taking the first predetermined number of clusters corresponding to the peak value that meets a preset condition as the actual number of clusters, wherein the preset condition is that a difference between the peak value and the maximum value is less than a preset value;
• clustering the service data in the second data set according to the actual number of clusters.

8:

• A classification device for service data comprising:
• a service data acquisition module for acquiring service data, each of which includes a plurality of service indicators;
• a service data extraction module for extracting the service data whose category attribute meets a preset condition from the service data according to an set extraction rule to form a first data set, wherein, the extraction rule is set according to part of the service indicators;
• a service data clustering module for taking the service data that are not extracted as a second data set, and clustering the service data in the second data set;
• a service data classification module for determining a classification result of the service data according to the first data set and a clustering result of the second data set.

9:

• The classification device according to claim 8, wherein the service data extraction module comprises at least one of the following units:
• a first extraction unit for extracting the service data, of which a certain service indicator data is an outlier according to a set threshold to form a first data set;
• a second extraction unit for extracting, according to a logical operation result of a plurality of service indicators of the service data, the service data, of which the logical operation result is an outlier to form a first data set;
• a third extraction unit for extracting, according to a data distribution of a certain service indicator, service data of which the certain service indicator data is an outlier, to form a first data set.

11:

• The classification device according to claim 10, wherein
• the service data classification module comprises an average feature acquisition unit and a merging unit;
• the average feature acquisition unit is used for determining an average feature value of service indicators involved in the extraction rule in the first data set and the second data set, wherein the average feature value is an average value or a median of the service indicators involved in the extraction rule of each class;
• the merging unit is used for merging the classes in the first data set with the classes in the second data set having the closest category attributes to the classes in the first data set to obtain the classification result of the service data;
• or,
• the service data classification module is used for merging the classes in the first data set with the classes in the second data set having the closest category attributes to the classes in the first data set to obtain the classification result of the service data, and making the difference in the number of the service data between each service data classification in the classification result conforms to a preset range.

13:

• The classification device according to claim 8, further comprising a pre-processing module which comprises:
• an outlier processing unit for selecting service indictor data that are outliers from the second data set, and assigning a service indicator demarcation value used to determine whether the service indicator data are outliers to the service indicator data that are outliers;
• and/or,
• a null value processing unit for calculating a mean value of all non-null data of a certain service indicator of the second data set, and assigning the mean value to null values of the service indicator in the second data set.

14:

• The classification device according to claim 8, wherein the service data clustering module comprises:
• a pre-clustering unit for pre-clustering the service data in the second data set according to each predetermined number of clusters, respectively;
• a profile coefficient calculation unit for calculating a profile coefficient of a pre-clustering result corresponding to each predetermined number of clusters;
• an actual cluster number determining unit for ranking the corresponding profile coefficients in an ascending order of the predetermined number of clusters, and acquiring several peak values in the profile coefficients and determining a maximum value therefrom, and taking the first predetermined number of clusters corresponding to the peak value that meets a preset condition as the actual number of clusters, wherein the preset condition is that a difference between the peak value and the maximum value is less than a preset value;
• an actual clustering unit for clustering the service data in the second data set according to the actual number of clusters.

15:

• A classification device for service data comprising:
• a memory;
• and
• a processor coupled to the memory, which is configured to execute the classification method for service data according to claim 1 on a basis of instructions stored in the memory.