DATA PROCESSING METHOD AND APPARATUS

• US 20200134361A1
• Filed: 12/27/2019
• Published: 04/30/2020
• Est. Priority Date: 06/30/2017
• Status: Active Grant

• Current Assignee: Huawei Technologies Co., Ltd. (Huawei Investment & Holding Co., Ltd.)
• Original Assignee: Huawei Technologies Co., Ltd. (Huawei Investment & Holding Co., Ltd.)
• Inventors: TANG, Ruiming, GUO, Huifeng, LI, Zhenguo, HE, Xiuqiang
• Granted Patent: US 11,334,758 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 18/2133 based on naturality criteria, e.g. with non-negative factorisation or negative correlation
• CPC Class Codes: G06F 18/2321 using statistics or function optimisation, e.g. modelling of probability density functions
• CPC Class Codes: G06F 18/245 relating to the decision surface
• CPC Class Codes: G06F 18/2451 linear, e.g. hyperplane
• CPC Class Codes: G06F 18/2453 non-linear, e.g. polynomial classifier
• CPC Class Codes: G06Q 10/04 Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202

Abstract

The method includes: obtaining a plurality of pieces of feature data; automatically performing two different types of nonlinear combination processing operations on the plurality of pieces of feature data to obtain two groups of processed data, where the two groups of processed data include a group of higher-order data and a group of lower-order data, the higher-order data is related to a nonlinear combination of m pieces of feature data in the plurality of pieces of feature data, and the lower-order data is related to a nonlinear combination of n pieces of feature data in the plurality of pieces of feature data, where m≥3, and m>n≥2; and determining prediction data based on a plurality of pieces of target data, where the plurality of pieces of target data include the two groups of processed data.

Claims

1. A data processing method, comprising:

obtaining a plurality of pieces of feature data;

automatically performing two different types of nonlinear combination processing operations on the plurality of pieces of feature data to obtain two groups of processed data, wherein the two groups of processed data comprise a group of higher-order data and a group of lower-order data, wherein the higher-order data is related to a nonlinear combination of m pieces of feature data in the plurality of pieces of feature data, and wherein the lower-order data is related to a nonlinear combination of n pieces of feature data in the plurality of pieces of feature data, wherein m≥

3, and m>

n≥

2; and

determining prediction data based on a plurality of pieces of target data, wherein the plurality of pieces of target data comprise the two groups of processed data.

2. The method according to claim 1, wherein automatically performing the two different types of nonlinear combination processing operations on the plurality of pieces of feature data to obtain two groups of processed data comprises:

automatically performing a higher-order nonlinear combination processing operation on the plurality of pieces of feature data to obtain the group of higher-order data; and

automatically performing a lower-order nonlinear combination processing operation on the plurality of pieces of feature data to obtain the group of lower-order data.

3. The method according to claim 1, wherein after obtaining the plurality of pieces of feature data, the method further comprises:

automatically performing a linear combination processing operation on the plurality of pieces of feature data to obtain a group of first auxiliary data, wherein the group of first auxiliary data is related to a linear combination of at least two pieces of feature data in the plurality of pieces of feature data, and wherein the plurality of pieces of target data further comprise at least one piece of the group of first auxiliary data.

4. The method according to claim 3,

wherein automatically performing the lower-order nonlinear combination processing operation on the plurality of pieces of feature data to obtain the group of lower-order data comprise automatically performing a lower-order nonlinear combination processing operation on the plurality of pieces of feature data by using a factorization machine (FM) based processing module or a field-aware factorization machine FFM-based processing module, to obtain the group of lower-order data;

and

wherein automatically performing the linear combination processing operation on the plurality of pieces of feature data to obtain the group of first auxiliary data comprises;

automatically performing a linear combination processing operation on the plurality of pieces of feature data by using the FM-based processing module or the FFM-based processing module, to obtain the group of first auxiliary data.

5. The method according to claim 1, wherein before determining the prediction data based on the plurality of pieces of target data, the method further comprises:

determining x pieces of empirical feature data with reference to both expert experience and the plurality of pieces of feature data, wherein x≥

1, and wherein the x pieces of empirical feature data comprise;

one or more of data obtained from the plurality of pieces of feature data with reference to the expert experience or data obtained by performing a nonlinear combination processing operation on at least two pieces of feature data in the plurality of pieces of feature data with reference to the expert experience; and

automatically processing the x pieces of empirical feature data to obtain a group of second auxiliary data, wherein the plurality of pieces of target data further comprise the group of second auxiliary data; and

wherein when x=1, the group of second auxiliary data is related to one piece of empirical feature data;

or when x>

1, the group of second auxiliary data is related to one or more of a first combination of data or a second combination of data, wherein the first combination of data is a linear combination of at least two pieces of empirical feature data in the x pieces of empirical feature data, and wherein the second combination of data is a nonlinear combination of at least two pieces of empirical feature data in the x pieces of empirical feature data.

6. The method according to claim 1, wherein the plurality of pieces of feature data are a plurality of pieces of sparse feature data;

and

after obtaining the plurality of pieces of feature data, the method comprises;

converting the plurality of pieces of sparse feature data into a plurality of pieces of dense feature data; and

wherein automatically performing the two different types of nonlinear combination processing operations on the plurality of pieces of feature data to obtain two groups of processed data comprises;

automatically performing two different types of nonlinear combination processing operations on the plurality of pieces of dense feature data to the obtain two groups of processed data.

19. A non-statutory computer readable storage medium, wherein the computer readable storage medium stores a computer program, and when the program is executed by a processor, the data processing method according to claim 1 is implemented.

7. A data processing apparatus, comprising:

an obtaining unit, configured to obtain a plurality of pieces of feature data;

a nonlinear processing unit, configured to automatically perform two different types of nonlinear combination processing operations on the plurality of pieces of feature data to obtain two groups of processed data, wherein the two groups of processed data comprise a group of higher-order data and a group of lower-order data, wherein the higher-order data is related to a nonlinear combination of m pieces of feature data in the plurality of pieces of feature data, and wherein the lower-order data is related to a nonlinear combination of n pieces of feature data in the plurality of pieces of feature data, wherein m≥

3, and m>

n≥

2; and

a predicting unit, configured to determine prediction data based on a plurality of pieces of target data, wherein the plurality of pieces of target data comprise the two groups of processed data.

8. The data processing apparatus according to claim 7, further comprising:

a higher-order processing subunit, configured to automatically perform a higher-order nonlinear combination processing operation on the plurality of pieces of feature data to obtain the group of higher-order data; and

a lower-order processing subunit, configured to automatically perform a lower-order nonlinear combination processing operation on the plurality of pieces of feature data to obtain the group of lower-order data.

9. The data processing apparatus according to claim 7, further comprising:

a linear processing unit, configured to automatically perform a linear combination processing operation on the plurality of pieces of feature data to obtain a group of first auxiliary data, wherein the group of first auxiliary data is related to a linear combination of at least two pieces of feature data in the plurality of pieces of feature data, and wherein the plurality of pieces of target data further comprise at least one piece of the group of first auxiliary data.

10. The data processing apparatus according to claim 9, further comprising:

a factorization machine FM-based processing module or a field-aware factorization machine FFM-based processing module;

wherein the FM-based processing module comprises a lower-order processing subunit and the linear processing unit; and

wherein the FFM-based processing module comprises a lower-order processing subunit and the linear processing unit.

11. The data processing apparatus according to claim 7, further comprising:

a determining unit, configured to determine x pieces of empirical feature data with reference to both expert experience and the plurality of pieces of feature data, wherein x≥

1, and wherein the x pieces of empirical feature data comprise;

one or more data obtained from the plurality of pieces of feature data with reference to the expert experience or data obtained by performing a nonlinear combination processing on at least two pieces of feature data in the plurality of pieces of feature data with reference to the expert experience; and

an empirical processing unit, configured to automatically process the x pieces of empirical feature data to obtain a group of second auxiliary data, wherein the plurality of pieces of target data further comprise the group of second auxiliary data; and

wherein when x=1, the group of second auxiliary data is related to one piece of empirical feature data;

or wherein when x>

1, the group of second auxiliary data is related to one or more of a first combination of data or a second combination of data, wherein the first combination of data is a linear combination of at least two pieces of empirical feature data in the x pieces of empirical feature data, and wherein the second combination of data is a nonlinear combination of at least two pieces of empirical feature data in the x pieces of empirical feature data.

12. The data processing apparatus according to claim 7, wherein the plurality of pieces of feature data are a plurality of pieces of sparse feature data, and the data processing apparatus further comprises:

a conversion unit, configured to convert the plurality of pieces of sparse feature data into a plurality of pieces of dense feature data, wherein

the nonlinear processing unit is further configured to;

automatically perform two different types of nonlinear combination processing operations on the plurality of pieces of dense feature data to the obtain two groups of processed data.

13. A data processing apparatus, further comprising:

a memory configured to store instructions; and

a processor coupled to the memory and configured to execute the instructions, which cause the processor to be configured to;

obtain a plurality of pieces of feature data;

automatically perform two different types of nonlinear combination processing operations on the plurality of pieces of feature data to obtain two groups of processed data, wherein the two groups of processed data comprise a group of higher-order data and a group of lower-order data, wherein the higher-order data is related to a nonlinear combination of m pieces of feature data in the plurality of pieces of feature data, and wherein the lower-order data is related to a nonlinear combination of n pieces of feature data in the plurality of pieces of feature data, wherein m≥

3, and m>

n≥

2; and

determine prediction data based on a plurality of pieces of target data, wherein the plurality of pieces of target data comprise the two groups of processed data.

14. The apparatus according to claim 13, the processor further executing the instructions to:

automatically perform a higher-order nonlinear combination processing operations on the plurality of pieces of feature data to obtain the group of higher-order data; and

automatically perform a lower-order nonlinear combination processing operations on the plurality of pieces of feature data to obtain the group of lower-order data.

15. The apparatus according to claim 13, the processor further executing the instructions to:

automatically perform a linear combination processing operations on the plurality of pieces of feature data to obtain a group of first auxiliary data, wherein the group of first auxiliary data is related to a linear combination of at least two pieces of feature data in the plurality of pieces of feature data, and wherein the plurality of pieces of target data further comprise at least one piece of the first auxiliary data.

16. The apparatus according to claim 15,

wherein automatically performing the lower-order nonlinear combination processing operations on the plurality of pieces of feature data to obtain the group of lower-order data comprises:

automatically a performing lower-order nonlinear combination processing operation on the plurality of pieces of feature data by using a factorization machine (FM) based processing module or a field-aware factorization machine FFM-based processing module, to obtain the group of lower-order data; and

wherein automatically performing the linear combination processing operation on the plurality of pieces of feature data to obtain a group of first auxiliary data comprises;

automatically performing a linear combination processing operation on the plurality of pieces of feature data by using the FM-based processing module or the FFM-based processing module, to obtain the group of first auxiliary data.

17. The apparatus according to claim 13, the processor further executing the instructions to:

determine x pieces of empirical feature data with reference to both expert experience and the plurality of pieces of feature data, wherein x≥

1, and wherein the x pieces of empirical feature data comprise;

one or more of data obtained from the plurality of pieces of feature data with reference to the expert experience or data obtained by performing a nonlinear combination processing operation on at least two pieces of feature data in the plurality of pieces of feature data with reference to the expert experience; and

automatically process the x pieces of empirical feature data to obtain a group of second auxiliary data, wherein the plurality of pieces of target data further comprise the group of second auxiliary data; and

wherein when x=1, the group of second auxiliary data is related to one piece of empirical feature data;

or when x>

1, the group of second auxiliary data is related to one or more of a first data combination or a second data combination, wherein the first data combination is a linear combination of at least two pieces of empirical feature data in the x pieces of empirical feature data, and wherein the second combination of data is a nonlinear combination of at least two pieces of empirical feature data in the x pieces of empirical feature data.

18. The apparatus according to claim 13, wherein the plurality of pieces of feature data are a plurality of pieces of sparse feature data;

and

after obtaining the plurality of pieces of feature data, the processor further executing the instructions to;

convert the plurality of pieces of sparse feature data into a plurality of pieces of dense feature data; and

wherein automatically performing the two different types of nonlinear combination processing operations on the plurality of pieces of feature data to obtain two groups of processed data comprises;

automatically perform two different types of nonlinear combination processing operations on the plurality of pieces of dense feature data to obtain the two groups of processed data.