Skip to content

Commit

Permalink
Create Conventional statistics vs machine learning
Browse files Browse the repository at this point in the history
  • Loading branch information
MoganaD committed Nov 1, 2020
1 parent 2cbfef8 commit f504f82
Showing 1 changed file with 19 additions and 0 deletions.
19 changes: 19 additions & 0 deletions Conventional statistics vs machine learning
Original file line number Diff line number Diff line change
@@ -0,0 +1,19 @@
The similar concepts between conventional statistics and machine learning


1. To determine the quality of data:

a) effect size or power analysis (Cox regression model) is done using conventional statistics.
The method used is residual analysis, if necessary standardized residual error will be performed using linear regression.

b) In machine learning, model evaluation is performed using different algorithms
The R codes for model evaluation is in the file (model evaluation)

2. To determine the significance of factors in the dataset relating to the target variable (dependent variable):

a) Chi sqauare test/ logistic regression is done in conventional statistics to select the significant variable

b) Variable importance is done in machine learning to determine the rank of variables based on importance score
(refer to variable importance R code)

3. Survival analysis R codes are under the survival analysis file.

0 comments on commit f504f82

Please sign in to comment.