LG Aimers 4기 B2B 고객 행동 예측 방법론

LG Aimers: AI전문가과정 4차

Module 8. 『B2B 고객데이터 기반 예측 단서 스코어링 모델』

ㅇ 교수 : KAIST 박성혁 교수 
ㅇ 학습목표     
본 모듈은 B2B 고객데이터 기반 예측 단서 스코어링 모델에 대해 학습합니다. 
고객의 행동을 예측할 수 있는 방법론과, 추천 시스템에 기반한 고객과 상품을 스코어링하는 방법, 의사결정나무 및 로지스틱 회귀 분석 기반의 고객과 상품을 스코어링하는 방법에 대해 소개합니다.

 

B2B 고객데이터 기반 예측 단서 스코어링 모델

Part 1. B2B 고객 행동 예측 방법론

-B2B 고객 행동 예측 방법론
:“targeting” 
Who? (Buyer/Churn)
ㄴBinary classification
 Logistic regression
 ANN
 Decision tree
 K-nearest neighbor
 SVM
 RFM (CRM)

:“product matching”
What? (Product)
ㄴrecommendations
 Content-based filtering
 Collaborative filtering

:“right time” 
When?
 Purchase interval 

:“expected revenue”
How much?
 Demand forecasting

-B2B vs B2C

 

B2B vs B2C

[reference] https://atonce.com/blog/thedifference-between-b2b-and-b2c-marketing

 

-Data Analytics
• Descriptive Analytics
 Describes the past status of the domain of interest using a variety of tools through techniques such as reporting, data visualization, dashboards, and scorecards

• Predictive Analytics
 Applies statistical and computational methods and models to data regarding past and current events to predict what might happen in the future

• Prescriptive Analytics
 Uses results of predictive analytics along with optimization and simulation tools to recommend actions that will lead to a desired outcome

-데이터 소스 측면에서의 세 가지 발전 단계

• Business Intelligence & Analytics 1.0
 Focus on structured quantitative data largely from relational databases

• Business Intelligence & Analytics 2.0
 Include data from the Web (Web interaction logs, customer reviews, social media)

• Business Intelligence & Analytics 3.0
 Include data from mobile devices, (location, sensors, etc.) and Internet of Things

 

 

[ +최신 트렌드 ]
딥러닝 기반 인공지능 기술은 영상 및 음성 정보를 손쉽게 처리해주기 때문에 더 풍부한 데이터를 기반으로 Analytics를 할 수 있게 한다.