九十七學年度下學期 類神經網路 研究計畫書

一、           研究計畫中英文摘要:請就本計畫要點作一概述,並依本計畫性質自訂關鍵詞。(五百字以內)

 

Abstract

 

Customer Relationship Management (CRM) is a key element of modern marketing strategies. The KDD Cup 2009 offers the opportunity to work on large marketing databases from the French Telecom company Orange to predict the propensity of customers to switch provider (churn), buy new products or services (appetency), or buy upgrades or add-ons proposed to them to make the sale more profitable (up-selling). The goal of KDD Cup 2009 is training Neural Networks by Customer Relationship Management (CRM) dataset which include more than 50000 data and 15000 variables, and to predict the churn, appetency and up-selling probability of customers.

 

Task Description

 

The task is to estimate the churn, appetency and up-selling probability of customers, hence there are three target values to be predicted.   

1.   Churn (wikipedia definition): Churn rate is also sometimes called attrition rate. It is one of two primary factors that determine the steady-state level of customers a business will support. In its broadest sense, churn rate is a measure of the number of individuals or items moving into or out of a collection over a specific period of time. The term is used in many contexts, but is most widely applied in business with respect to a contractual customer base. For instance, it is an important factor for any business with a subscriber-based service model, including mobile telephone networks and pay TV operators. The term is also used to refer to participant turnover in peer-to-peer networks.

2.  Appetency: In our context, the appetency is the propensity to buy a service or a product.

3.  Up-selling (wikipedia definition): Up-selling is a sales technique whereby a salesman attempts to have the customer purchase more expensive items, upgrades, or other add-ons in an attempt to make a more profitable sale. Up-selling usually involves marketing more profitable services or products, but up-selling can also be simply exposing the customer to other options he or she may not have considered previously. Up-selling can imply selling something additional, or selling something that is more profitable or otherwise preferable for the seller instead of the original sale.

 

二、研究計畫內容:

 

() 1.Background

 

CRM (Customer Relationship Management), This concept was first proposed by Gartner Group, and it beginning popularity the enterprise e-commerce in the recent. The main meaning of CRM is the customer detailed information on in-depth analysis to improve customer satisfaction, thereby enhancing the competitiveness of enterprises as a means, it mainly includes the following aspects (the 7P):

1. Customer Profile Analysis (Profiling), including the level of customer risk, hobbies, habits.                                                                  

2. Analysis of customer loyalty (Persistency) refers to a product to customers or businesses faithfully the extent of persistence, and other changes.                

3. Analysis of customer profitability (Profitability) by means of different consumer products clients the edge of profits, the amount of gross profit, net profit, etc.

 4. Performance analysis of client (Performance) by referring to different customers according to the types of consumer products, channels and sales division of the location of the sales target.

 5. Future analysis of customer (Prospecting), including the number of customers, types of the future development trend of the means for customers.

6. Customer Product analysis (Product), including product design, relevance, such as supply

   chain.

7. Analysis of customer promotion (Promotion), including advertising, promotion and other publicity activities.

It is not just the software, and it is the methodology, software and IT capacity-General, is the business strategy.

 

() 2.Research goal

 

Because is getting progressive more and more along with the time, so the people purchase commodity changing fastidious about conveniently. Therefore bosses in order to must gain a bigger profit, they have to understand the custom and purchase rule of customers. The plans will customer's custom to doing the statistical and analysis and classifies. So it will understand that customers' habit, then, the store may according to customer's rule to make the adjustment, and then it is not only facilitated the customer, but also let the company earn a bigger profit. So this plain will make win-win result.

 

 

() 1.Method

 

 Method flow

 

 

 Step1First data will do preprocess make some noise elimination

 Step2From training data extraction a few data to do analysis

 Step3Select useful attribute to do analysis from training data

 Step4Using Matlab to analysis

 Step5Result analysis

 

 In step1 I will do preprocess, because they have a lot of blank or without value in the attribute, so I have to do preprocess and let the blank fill out 0 in training data. And in step2 I will extraction 1000 data and to select useful attribute because in the training data is too huge could cause I can not use Matlab to work. And I use Matlab Network type selecting Feed forward backprop, and training function selecting trainlm, and performance function selecting MSE, and finally it run result use Matlab.

 

() 2.Problem

 

1. Training data is too huge, let Matlab could not run

2. How many should you extraction data from training data?

3. Which attribute should you select?

4. The accuracy is too low?

 

() 2.Improve method

 

1. Using a few training data, or use high-level computer or tool platform

2. Because we do not know to select how many training data, so we can to do choosing the complete

   training data to divide into several parts.

3. It is only can try and try or to choose more attribute.

4. To improve the preprocess, because data could also have a lot of noise after preprocess.