Potential customer level acquisition method based on BBD or/and RF model and system thereof

A model acquisition and model technology, applied in marketing and other directions, can solve problems such as error-prone, limited data volume, and heavy analysis workload, and achieve the effect of ensuring effectiveness and eliminating artificial influence

Inactive Publication Date: 2015-11-25
CHEZHI HULIAN BEIJING SCI & TECH CO LTD
View PDF4 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] Most of the data is based on the database, so the amount of data that can be processed is limited, and a large amount of data analysis is often required to gain an in-depth understanding of user behavior; manual user behavior analysis is heavy and error-prone; manual use of prior knowledge Determining the weight of different user behaviors and making weight combinations, resulting in more manual interference, and the results cannot be measured, and the effect cannot be guaranteed

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
  • Potential customer level acquisition method based on BBD or/and RF model and system thereof
  • Potential customer level acquisition method based on BBD or/and RF model and system thereof
  • Potential customer level acquisition method based on BBD or/and RF model and system thereof

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0076] Such as figure 1 As shown, the embodiment of the present invention provides a method for obtaining the potential customer level based on the BBD model, including the following steps:

[0077] S1, constructing a BBD model, the form of the BBD model is as follows:

[0078]

[0079] In the formula,

[0080] C: the number of times the user placed an order in history,

[0081] I: the number of historical browsing behaviors of users,

[0082] Ctr: the user's potential order probability,

[0083] α, β: parameters of the BBD model;

[0084] S2, adopt the MLE method to establish the following formula:

[0085]

[0086] In the formula,

[0087] n: the number of samples;

[0088] S3, the loglikelihood function of complete data is obtained after taking ln from the formula in S2;

[0089] S4, calculating the mathematical expectation of the loglikelihood function of the complete data;

[0090] S5. Obtain an iterative formula of the following form according to the mathe...

Embodiment 2

[0121] Such as figure 2 As shown, the embodiment of the present invention provides a system for obtaining potential customer levels based on the BBD model, including:

[0122] Data platform: including feature data layer and portrait data layer. The feature data layer includes: historical behavior data of website users, behavior feature data mined from behavior data and feature data generated from existing user portrait data; portrait data layer Including: user portrait data obtained through model learning from the feature data;

[0123] Computing platform: including a model layer, an algorithm layer and a calculation layer, the model layer includes: BBD model; the algorithm layer includes: SGD / GD, LBFGS and CD optimization algorithms; the calculation layer reads the data platform The feature data is trained using the optimization algorithm provided by the algorithm layer to obtain the BBD model parameters;

[0124] Application platform: read the BBD model obtained by the co...

Embodiment 3

[0128] Such as image 3 As shown, the embodiment of the present invention provides a method for obtaining potential customer levels based on the BBD model and the RF model, including the following steps:

[0129] Step 1, utilize machine learning model, construct the RF fusion model based on BBD model and RF model; Wherein, the method for solving BBD model obtains the method for potential visitor level based on BBD model described in any one of claim 1-6;

[0130] Step 2, apply the RF fusion model and output potential-level data.

[0131] Wherein, said step 1 may include the following steps:

[0132] Step 101, divide the sample set into training sample set 1, training sample set 2 and test sample set;

[0133] Step 102, using the training sample set one to solve the BBD model, and the method for solving the BBD model is as described in any one of claims 1-6, obtaining the BBD model and outputting the result;

[0134] Step 103, using the second training sample set to solve th...

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 potential customer level acquisition method based on a BBD or / and RF model and a system thereof, and relates to the technical field of potential data mining. The potential customer level is acquired by constructing and using the BBD or / and RF model, historical behavior data of users are fully utilized, and the behavior data of the users in multiple days (e.g. 120 days) are extracted by utilizing a big data platform so that user behaviors can be profoundly understood; besides, a characteristic engineering system is constructed, the behavior characteristics of the users, such as preferences, time duration and frequency, are automatically computed, a mode identification method is adopted, and the model is enabled to learn the law of the data and automatically confirm the weight relation between an independent variable and a dependent variable through reasonable assumption of probabilistic distribution of the data so that a problem of heavy workload of artificial user behavior analysis can be solved and artificial influence can also be completely eliminated. Meanwhile, the measurement index is clear so that effectiveness of the output potential customer level can be guaranteed.

Description

technical field [0001] The present invention relates to the technical field of potential data mining, in particular to a method and system for obtaining potential customer levels based on BBD and / or RF models. Background technique [0002] Every day, some potential car buyers will convert into actual car buyers with a high probability in a short period of time (for example, users who have ordered SL behaviors). Level, you can predict users in advance, so as to help manufacturers, dealers and websites to make corresponding operating strategies to improve or affect user experience or decision-making process. Therefore, the mining of potential customers (referred to as potential customers for short) is an important part of the car sales website UP. [0003] At present, the following methods are generally used for potential customer analysis: [0004] Sampling user behavior logs from the database, but due to the amount of data, the time period is short; manual analysis of user...

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): G06Q30/02
Inventor 张磊
Owner CHEZHI HULIAN BEIJING SCI & TECH CO LTD
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