Decision_Tree

Lab1 - Decision Tree

Introduction

Experimenting with decision tree based inductive classifier using early stopping, noise, post pruning, random forests.

• Please refer to Report.pdf for detailed analysis.
• Please refer to lab.pdf for about the project.

Directory Structure

---README
---lab.pdf
---Report.pdf
---code
	|
	|--decision.cpp
	|--imdbEr.txt
	|--imdb.vocab
	|--makefile
	|--script.sh
	|--script_skip1.sh
	|--selected-features-indices.txt
	|
	|--train
	|	|
	|	|--labeledBow.feat
	|
	|
	|--test
		|
		|--labeledBow.feat

Executing

Make the code using $make or $g++ decision.cpp -o decision

Exceute using ./decision filename expno polarity -- filename refers to the selected indices file -- refers to expno from 1-5 -- polarity is optional. 0 (default) = use most frequent words; 1 = use highest and lowest polarity words

You can use script to run all functionality using $./script.sh

To use previously complied selected-features-indices.txt and skip expno 1 use $./script_skip1.sh

Note: It script fails to exceute use $chmod 755 script.sh

Feel free to email at [email protected] for any clarification.

Developed by

Naman Goyal