Unlock instant, AI-driven research and patent intelligence for your innovation.

User behavior tracking data analysis method and system and storage medium

An analysis method and analysis system technology, applied in the analysis method, system and storage medium field of user behavior tracking data, can solve problems such as single data source, inability to perform effective analysis, and inability to take into account useful order data

Inactive Publication Date: 2019-10-18
携程旅游信息技术(上海)有限公司
View PDF6 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The use of order data for data mining and user identification is relatively single in terms of data sources, and useful data other than orders cannot be taken into account; users who forge their identities or forge device fingerprints cannot be effectively analyzed

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
  • User behavior tracking data analysis method and system and storage medium
  • User behavior tracking data analysis method and system and storage medium
  • User behavior tracking data analysis method and system and storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0050] This embodiment provides a method for analyzing user behavior tracking data, such as figure 1 As shown, the analysis method of user behavior tracking data includes the following steps:

[0051] S1. Collect user behavior tracking data of the two orders during the process of placing an order;

[0052] To analyze the UBT data corresponding to the order, it is first necessary to collect the behavior data of the user who placed the order after logging in to the website. After the user logs in to the website, the user behavior tracking data is complicated. For example: the user searches for products in the search box, clicks on one of the products, browses the pictures of the product, scrolls to the product reviews, browses the product reviews, Bookmarked the product, clicked on the relevant recommendations under the product, browsed the relevant recommendations, closed the relevant recommendations, returned to the previous product page to place an order, filled in the infor...

Embodiment 2

[0057] This embodiment is obtained by further refinement and extension on the basis of embodiment 1. Considering that user behavior tracking data is unstructured and cannot be directly used for similarity calculation, this embodiment performs a series of calculations on unstructured UBT data to obtain order similarity. Such as figure 2 As shown, the S2 step of the present embodiment includes the following 3 steps:

[0058] S21. Using the one-hot encoding structured user behavior tracking data to obtain a user behavior tracking matrix.

[0059] After the user logs in to the website, many user behaviors are added with posterior business labels to construct a complete x (feature) and y (label) of UBT data. After collecting the UBT data of the two orders, the next step is to UBT encode the UBT x data. Before using one-hot encoding (one-hot) to encode the behaviors in the above example, such as: login, browse, and search, you must first define how many user behaviors there are ...

Embodiment 3

[0071] In this embodiment, order similarity is used to determine order attributes. Such as Figure 5 As shown, the analysis method of user time series behavior data in this embodiment includes the following three steps:

[0072] S301. Collect user behavior tracking data of several orders, and use step S2 to calculate the similarity between two orders;

[0073] Using the UBT data corresponding to a large number of orders, use step S2 to calculate the similarity between pairs of orders. See Example 2 for the detailed calculation process.

[0074] S302. Based on the similarity between pairs of orders, use the K-centroid algorithm to cluster the orders to obtain several classes;

[0075] With the distance between orders, k-medoids (K-central point algorithm) can be used to cluster a large number of orders, and cluster similar orders to form different classes.

[0076] S303. Analyze the data characteristics of each class. The data features include the features that indicate whe...

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 user behavior tracking data analysis method and system, and a storage medium. The analysis method comprises the following steps: S1, collecting user behavior tracking data oftwo orders in a user ordering process; and S2, calculating the similarity of the two orders by using the user behavior tracking data. According to the method, the user behavior tracking data of two orders in the user ordering process are collected, the similarity of the two orders is calculated by utilizing the data. A large number of orders can be clustered by utilizing the similarity of the orders, and then all the classes are analyzed. The techniques, characteristics, modes and the like of fraudulent orders are summarized from the classes, possible fraudulent orders or fraudulent users canbe found in the early stage, loss is stopped by follow-up processing, companies are powerfully protected, and fraudulent behaviors are attacked.

Description

technical field [0001] The present invention relates to the field of computer software, in particular to an analysis method, system and storage medium for User BehaviorTrace (UBT, User BehaviorTrace) data. Background technique [0002] The era of artificial intelligence and big data has arrived. Under this general trend, the risk control department needs to keep pace with the times and try to mine user data from more angles, so as to distinguish good users from potentially fraudulent and malicious ones. Refusal to achieve the purpose of protecting enterprises and combating black industries. [0003] In the past, data mining and user screening mostly focused on order data. The order data currently used mainly include user portraits (such as the transaction status of this user in a certain time window in the past, transparency of user information, whether they are blacklisted, etc.) , Product portrait (whether the origin and destination of the order belong to a high-risk area...

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
IPC IPC(8): G06Q30/06G06K9/62G06N3/04
CPCG06Q30/0609G06N3/045G06F18/23213G06F18/22
Inventor 陈嘉东霍文虎
Owner 携程旅游信息技术(上海)有限公司