Predicting types of records based on amount values of records

1. A non-transitory machine-readable medium storing a program executable by at least one processing unit of a device, the program comprising sets of instructions for:

querying a database for a subset of a plurality of records in the database, each record in the plurality of records comprising a first value for a first field and a second value for a second field;

normalizing the first value of the first field of each record in the subset of the plurality of records;

dividing the subset of the plurality of records into a plurality of groups of records based on the second values of the second field; and

generating a function for predicting a type of a particular record based on a value of a field of the particular record and the first values of the first field of records in the plurality of groups of records.

2. The non-transitory machine-readable medium of claim 1 , wherein generating the function comprises, for each group of records, calculating a descriptive statistic of the first values of the first field of records in the group of records,

wherein the function determines the type of the particular record by:

determining a descriptive statistic in the descriptive statistics calculated for the plurality of groups of records that is closest to the value of the field of the particular record and

determining the type of the particular record based on the second values of records in the group of records from which the determined descriptive statistic is calculated.

3. The non-transitory machine-readable medium of claim 2 , wherein the descriptive statistic is a mean value, a median value, or a mode value.

4. The non-transitory machine-readable medium of claim 1 , wherein generating the function comprises, for each group of records, calculating a probability density function of the first values of the first field of records in the group of records,

wherein the function determines the type of the particular record by:

calculating, for each group of records in the plurality of groups of records, a probability value using the value of the field of the particular record as input to the probability density function calculated for the group of records,

determining a highest probability value in the probability values calculated for the plurality of groups of records, and

determining the type of the particular record based on the second values of records in the group of records from which the probability density function of the determined highest probability value is calculated.

5. The non-transitory machine-readable medium of claim 4 , wherein calculating the probability density function for each group of records comprises using a kernel density estimation algorithm.

6. The non-transitory machine-readable medium of claim 1 , wherein each record in the plurality of records further comprises a third value for a third field, wherein normalizing the first value for the first field of each record in the subset of the plurality of records comprises:

calculating a quotient by dividing the first value for the first field by the third value of the third field; and

using the quotient as the normalized first value for the first field.

7. The non-transitory machine-readable medium of claim 1 , wherein dividing the subset of the plurality of records into the plurality of groups of records based on the second values of the second field comprises dividing the subset of the plurality of records into the plurality of groups of records so that records in each group of records have the same second values for the second field.

8. A method comprising:

querying a database for a subset of a plurality of records in the database, each record in the plurality of records comprising a first value for a first field and a second value for a second field;

normalizing the first value of the first field of each record in the subset of the plurality of records;

dividing the subset of the plurality of records into a plurality of groups of records based on the second values of the second field; and

generating a function for predicting a type of a particular record based on a value of a field of the particular record and the first values of the first field of records in the plurality of groups of records.

9. The method of claim 8 , wherein generating the function comprises, for each group of records, calculating a descriptive statistic of the first values of the first field of records in the group of records,

wherein the function determines the type of the particular record by:

determining a descriptive statistic in the descriptive statistics calculated for the plurality of groups of records that is closest to the value of the field of the particular record and

determining the type of the particular record based on the second values of records in the group of records from which the determined descriptive statistic is calculated.

10. The method of claim 9 , wherein the descriptive statistic is a mean value, a median value, or a mode value.

11. The method of claim 8 , wherein generating the function comprises, for each group of records, calculating a probability density function of the first values of the first field of records in the group of records,

wherein the function determines the type of the particular record by:

calculating, for each group of records in the plurality of groups of records, a probability value using the value of the field of the particular record as input to the probability density function calculated for the group of records,

determining a highest probability value in the probability values calculated for the plurality of groups of records, and

determining the type of the particular record based on the second values of records in the group of records from which the probability density function of the determined highest probability value is calculated.

12. The method of claim 11 , wherein calculating the probability density function for each group of records comprises using a kernel density estimation algorithm.

13. The method of claim 8 , wherein each record in the plurality of records further comprises a third value for a third field, wherein normalizing the first value for the first field of each record in the subset of the plurality of records comprises:

calculating a quotient by dividing the first value for the first field by the third value of the third field; and

using the quotient as the normalized first value for the first field.

14. The method of claim 8 , wherein dividing the subset of the plurality of records into the plurality of groups of records based on the second values of the second field comprises dividing the subset of the plurality of records into the plurality of groups of records so that records in each group of records have the same second values for the second field.

15. A system comprising:

a set of processing units; and

a non-transitory machine-readable medium storing instructions that when executed by at least one processing unit in the set of processing units cause the at least one processing unit to:

query a database for a subset of a plurality of records in the database, each record in the plurality of records comprising a first value for a first field and a second value for a second field;

normalize the first value of the first field of each record in the subset of the plurality of records;

divide the subset of the plurality of records into a plurality of groups of records based on the second values of the second field; and

generate a function for predicting a type of a particular record based on a value of a field of the particular record and the first values of the first field of records in the plurality of groups of records.

16. The system of claim 15 , wherein generating the function comprises, for each group of records, calculating a descriptive statistic of the first values of the first field of records in the group of records,

wherein the function determines the type of the particular record by:

determining a descriptive statistic in the descriptive statistics calculated for the plurality of groups of records that is closest to the value of the field of the particular record and

determining the type of the particular record based on the second values of records in the group of records from which the determined descriptive statistic is calculated.

17. The system of claim 16 , wherein the descriptive statistic is a mean value, a median value, or a mode value.

18. The system of claim 15 , wherein generating the function comprises, for each group of records, calculating a probability density function of the first values of the first field of records in the group of records,

wherein the function determines the type of the particular record by:

calculating, for each group of records in the plurality of groups of records, a probability value using the value of the field of the particular record as input to the probability density function calculated for the group of records,

determining a highest probability value in the plurality of probability values calculated for the plurality of groups of records, and

determining the type of the particular record based on the second values of records in the group of records from which the probability density function of the determined highest probability value is calculated.

19. The system of claim 18 , wherein calculating the probability density function for each group of records comprises using a kernel density estimation algorithm.

20. The system of claim 15 , wherein each record in the plurality of records further comprises a third value for a third field, wherein normalizing the first value for the first field of each record in the subset of the plurality of records comprises:

calculating a quotient by dividing the first value for the first field by the third value of the third field; and

using the quotient as the normalized first value for the first field.