Data-driven method and apparatus for handling user inquiries using collected data

Patent metadata

Field	Value
Publication number	US-11928617-B2
Application number	US-202218045801-A
Country	US
Kind code	B2
Filing date	Oct 11, 2022
Priority date	Jan 8, 2016
Publication date	Mar 12, 2024
Grant date	Mar 12, 2024

Abstract

The present disclosure provides data-driven methods and apparatuses for predicting user inquiries. One exemplary method includes: collecting user behavior data and pre-processing the user behavior data when a user inquiry is received; extracting candidate user behavior data that is contributive to the user inquiry from the pre-processed user behavior data; screening the candidate user behavior data based on a set target behavior data set, and selecting candidate user behavior data that is contained in the target behavior data set; inputting the screened candidate user behavior data into a trained classifier model; and predicting an inquiry category to which the user inquiry belongs. One exemplary apparatus includes a pre-processing module, an extraction module, and a prediction module. The method and the apparatus embodiments of the present disclosure can improve the efficiency and accuracy of the prediction.

First claim

The invention claimed is:

1. A method comprising:
collecting, by a device having one or more processors, user behavior data when a user inquiry is received;
extracting user behavior data associated with the user inquiry from the collected user behavior data as candidate user behavior data;
selecting, from the candidate user behavior data, candidate user behavior data that is contained in a target behavior data set;
inputting, by the device, the selected candidate user behavior data into a classifier model that is a neural network model trained based on training data; and
predicting an inquiry category associated with the user inquiry using the classifier model based on the inputted selected candidate user behavior data,
wherein the classifier model is trained by:
collecting a plurality of training user inquiries and training user behavior data corresponding thereto;
extracting user behavior data associated with each of the plurality of training user inquiries from the collected training user behavior data as training candidate user behavior data;
scoring, by using a data-driven method, the training candidate user behavior data corresponding to each of the plurality of training user inquiries;
selecting, from the scored training candidate user behavior data, target behavior data based on a set condition; and
obtaining the classifier model by training, based on the plurality of training user inquiries and the target behavior data.

2. The method of claim 1 , wherein extracting user behavior data associated with the user inquiry from the collected user behavior data as candidate user behavior data uses a windowing and truncation process comprising:
extracting user behavior data in a period of time prior to the user inquiry.

3. The method of claim 1 , wherein before extracting user behavior data associated with each of the plurality of training user inquiries from the collected training user behavior data as training candidate user behavior data, the collected training user behavior data is pre-processed by removing user behavior data having a frequency of occurrence lower than a set threshold.

4. The method of claim 1 , wherein before extracting user behavior data associated with each of the plurality of training user inquiries from the collected training user behavior data as training candidate user behavior data, the collected training user behavior data is pre-processed by: digitally identifying the collected user behavior data.

5. The method of claim 1 , wherein before obtaining the classifier model by training, the target behavior data is digitally identified.

6. The method of claim 1 , wherein before obtaining the classifier model by training, vectorization is performed on the target behavior data.

7. An apparatus, comprising:
a memory storing a set of instructions; and
a processor configured to execute the set of instructions to cause the apparatus to perform:
collecting user behavior data when a user inquiry is received;
extracting user behavior data associated with the user inquiry from the pre-processed user behavior data as candidate user behavior data;
selecting, from the candidate user behavior data, candidate user behavior data that is contained in a target behavior data set;
inputting the selected candidate user behavior data into a classifier model that is a neural network model trained based on training data; and
predicting an inquiry category associated with the user inquiry using the classifier model based on the inputted selected candidate user behavior data,

wherein the classifier model is trained by:
collecting a plurality of training user inquiries and training user behavior data corresponding thereto;
extracting user behavior data associated with each of the plurality of training user inquiries from the collected training user behavior data as training candidate user behavior data;
scoring, by using a data-driven method, the training candidate user behavior data corresponding to each of the plurality of training user inquiries;
selecting, from the scored training candidate user behavior data, target behavior data based on a set condition; and
obtaining the classifier model by training, based on the plurality of training user inquiries and the target behavior data.

8. The apparatus of claim 7 , wherein extracting user behavior data associated with the user inquiry from the collected user behavior data as candidate user behavior data uses a windowing and truncation process comprising:
extracting user behavior data in a period of time prior to the user inquiry.

9. The apparatus of claim 7 , wherein before extracting user behavior data associated with each of the plurality of training user inquiries from the collected training user behavior data as training candidate user behavior data, the collected training user behavior data comprises:
removing user behavior data having a frequency of occurrence lower than a set threshold.

10. The apparatus of claim 7 , wherein before extracting user behavior data associated with each of the plurality of training user inquiries from the collected training user behavior data as training candidate user behavior data, the collected training user behavior data is pre-processed by digitally identifying the collected user behavior data.

11. The apparatus of claim 7 , wherein before obtaining the classifier model by training, the target behavior data is digitally identified.

12. The method of claim 7 , wherein before obtaining the classifier model by training, vectorization is performed on the target behavior data.

13. A non-transitory computer readable medium that stores a set of instructions that is executable by at least one processor of a computer to cause the computer to perform a method, the method comprising:
extracting user behavior data associated with the user inquiry from collected user behavior data as candidate user behavior data;
selecting, from the candidate user behavior data, candidate user behavior data that is contained in a target behavior data set;
inputting the selected candidate user behavior data into a classifier model that is a neural network model trained based on training data; and
predicting an inquiry category associated with the user inquiry using the classifier model based on the inputted selected candidate user behavior data,
wherein the classifier model is trained by:
collecting a plurality of training user inquiries and training user behavior data corresponding thereto;
extracting user behavior data associated with each of the plurality of training user inquiries from the collected training user behavior data as training candidate user behavior data;
scoring, by using a data-driven method, the training candidate user behavior data corresponding to each of the plurality of training user inquiries;
selecting, from the scored training candidate user behavior data, target behavior data based on a set condition; and
obtaining the classifier model by training, based on the plurality of training user inquiries and the target behavior data.

14. The non-transitory computer readable medium of claim 13 , wherein extracting user behavior data associated with the user inquiry from the collected user behavior data as candidate user behavior data uses a windowing and truncation process comprising:
extracting user behavior data in a period of time prior to the user inquiry.

15. The non-transitory computer readable medium of claim 13 , wherein before extracting user behavior data associated with each of the plurality of training user inquiries from the collected training user behavior data as training c

Assignees

• Alibaba Group Holding Ltd

Inventors

• Xue Shaofei
• Zhang Jiaxing
• Cui Hengbin

Classifications

• G06Q10/04 Primary

  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) · CPC title

• G06F18/00

  Pattern recognition · CPC title

• G06N5/04

  Inference or reasoning models · CPC title

• G06N20/00

  Machine learning · CPC title

• G06Q30/016

  After-sales · CPC title