Predictive Analytics for Business 2

标签：set   ssi   rmi   stat   iat   drive   val   via   location   

Methodology Map:

Methodology map helps us to determine which method I am going to use to solve the business problem:

1. NON-predictive problems:
　　a. Geospatial analysis: this type of peoblem use location based data to drive your conclusions. (coordinates, distance, geographic location ect.)

　　b. Segmentation is to group data together.

　　c. Aggregation is the methodology that simply means calculating a value across a group or dimension and is commonly used in data analysis.

　　d. Descripitive statistics means to use statistics to describe the data.(Mean, Median, Mode, Standard Deviation, Interqurtile Range)

2. Data Rich vs. Data Poor:

　　How to determine Data rich or data poor?

　　Whether we have data on what we are trying to predict.

If data poor:

　　Setup an experiment  (A/B test)

3. Regression model vs Classfication model

4. Regression(numeric) model:

　　a. Continuous model.

　　b. Time-based model.

5.  Classfication(non-numeric) model:

　　a. Binary (yes or no)

　　b. Non-binary(mutiple opinions)

Predictive Analytics for Business 2

标签：set   ssi   rmi   stat   iat   drive   val   via   location   

原文地址：http://www.cnblogs.com/kingoscar/p/6056147.html