Consumption intention identification and prediction method based on consumption cause map

A prediction method and intent technology, applied in data processing applications, special data processing applications, other database retrieval, etc., can solve the problems of low accuracy of consumer intent recognition and prediction

Active Publication Date: 2020-12-25
HARBIN INST OF TECH
View PDF1 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0009] The purpose of the present invention is to solve the problem of low accuracy in the identification and prediction of consumption intentions in the prior art, and propose a method for identification and prediction of consumption intentions based on the consumer affairs map

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Consumption intention identification and prediction method based on consumption cause map
  • Consumption intention identification and prediction method based on consumption cause map
  • Consumption intention identification and prediction method based on consumption cause map

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0021] Specific implementation mode one: combine figure 1 This embodiment will be described. A method for identifying and predicting consumption intentions based on the consumer affairs map described in this embodiment is specifically implemented through the following steps:

[0022] Step 1. After fine-tuning the pre-trained BERT-Base model using the data marked with events, use the fine-tuned BERT-Base model to extract events from the narrative text;

[0023] Step 2. After fine-tuning the pre-trained BERT-Base model using the data marked with the event and the sequence relationship between events, use the fine-tuned BERT-Base model to make the sequence of the event pair composed of the events extracted in step 1. Identify the inheritance relationship and build a map of affairs;

[0024] Step 3. Construct a bipartite graph based on the events extracted in step 1 as weak supervision information;

[0025] Step 4. Using the bipartite graph embedding algorithm, combining the ev...

specific Embodiment approach 2

[0029] Embodiment 2: The difference between this embodiment and Embodiment 1 is that in the first step, the pre-trained BERT-Base model is fine-tuned using the data marked with events, and the specific process is as follows:

[0030] After inputting the labeled data into the pre-trained BERT-Base model, the fine-tuning process is carried out in two stages:

[0031] The first stage: set the learning rate to 1e-3, the training round is 2, the optimization algorithm used is the BertAdam algorithm, and the linear layer parameters of the pre-trained BERT-Base model are trained;

[0032] The second stage: set the learning rate to 3e-5, the training rounds to 10, the optimization algorithm used is the BertAdam algorithm, and the parameters of the linear layer and BERT layer of the pre-trained BERT-Base model are trained;

[0033] After completing the two-stage training, the fine-tuned BERT-Base model is obtained.

[0034] The annotation data used in this embodiment is the data after t...

specific Embodiment approach 3

[0036] Specific embodiment three: the difference between this embodiment and specific embodiment two is that in the second step, the pre-trained BERT-Base model is fine-tuned using the data marked with events and the sequential relationship between events. The process is:

[0037] After inputting the labeled data into the pre-trained BERT-Base model, the fine-tuning process is divided into three stages:

[0038] The first stage: train the parameters of the linear layer, the learning rate is set to 1e-3, the training round is 1, and the optimization algorithm used is the BertAdam algorithm;

[0039] The second stage: train the parameters of the linear layer and the event embedding layer, the learning rate is set to 5e-5, the training round is 3, and the optimization algorithm used is the BertAdam algorithm;

[0040] The third stage: train the parameters of all layers, the learning rate is set to 1e-5, the training round is 5, and the optimization algorithm used is the BertAdam...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a consumption intention identification and prediction method based on a consumption cause map, and belongs to the technical field of consumption intention identification and prediction. According to the invention, the problem of low consumption intention identification and prediction accuracy in the prior art is solved. According to the main technical scheme, the method comprises the steps: step 1, performing event extraction based on a pre-training model; step 2, extracting a relationship between events based on the pre-training model; step 3, constructing a bipartitegraph as weak supervision information by adopting an unsupervised method based on the comment corpus; step 4, constructing a consumption reason graph based on the bipartite graph weak supervision information and the annotation data; and step 5, training a homogeneity relationship attention model by using the training data, and judging a corresponding relationship between the event and the consumption intention by using the homogeneity relationship attention model. The method can be applied to consumption intention identification and prediction.

Description

technical field [0001] The invention belongs to the technical field of consumption intention identification and prediction, and in particular relates to a method for identification and prediction of consumption intention based on a graph of consumption affairs. Background technique [0002] Consumption intention refers to the willingness expressed by users to purchase products and services and other commercial consumption needs (Fu, B., and T. Liu."Weakly-supervised consumption intent detection in microblogs."Journal of Computational Information Systems 6.9(2013) :2423-2431.). Consumption intention recognition technology is to analyze and sort out the text generated by Internet users with the color of consumer demand or the user's own behavior data, and dig out the user's current or potential consumption demand in a specific way (Fu, B., and T. Liu ."Consumption intent recognition for social media: task, challenge and opportunity." Intell Comput Appl 5(2015):1-4.). Consump...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06Q30/02G06F16/951
CPCG06Q30/0201G06Q30/0202G06F16/951
Inventor 丁效秦兵刘挺石乾坤
Owner HARBIN INST OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap