aneesh404/Flair-detector

This repo contains code for a Flair Detector made as a submission for MIDAS Internship task

Jupyter Notebook

Reddit Flare Detection

This repository contains code for a machine learning model that detects the flair of the reddit post, given the url.

The repo is divided into the following structure:

Notebooks - Containing all the .pynb files related to the project.
Web App - containing all the files needed for running the web-app.
Model - containing the pickle file of the best performing model.

How to run locally

Open terminal and perform the following steps.

Clone the repository

git clone https://github.com/aneesh404/Flair-detector.git

Install the required libraries

pip3 install -r requirements.txt

Navigate to the folder containing flask code

cd web_app

Export the following variables in your terminal

export FLASK_APP=app.py && export FLASK_ENV=development

Run the app

flask run

Go to the address as shown after running the command above.

Results:

Features : Titles+Comments+Body+Url

	Model used	Accuracy
	`Random Forest Classifier`	68.983
	`naive_bayes`	57.657
	`linear_svm`	71.042
	`logistic regression`	70.656
	`Bag Of Words`	64.092

Features : Titles+Comments+Body

	Model used	Accuracy
	`Random Forest Classifier`	69.240
	`Naive Bayes`	55.083
	`Linear svm`	67.567
	`Logistic Regression`	69.112
	`Bag Of Words`	65.122

Features : Title+Body

	Model used	Accuracy
	`Random Forest Classifier`	65.637
	`Naive Bayes`	60.875
	`Linear svm`	66.795
	`Logistic Regression`	66.410
	`Bag Of Words`	64.052

Features : Title+Comments

	Model used	Accuracy
	`Random Forest Classifier`	60.231
	`Naive Bayes`	50.454
	`Linear svm`	63.191
	`Logistic Regression`	62.934
	`Bag Of Words`	61.122

Features : Title

	Model used	Accuracy
	`Random Forest Classifier`	54.697
	`Naive Bayes`	53.796
	`Linear svm`	53.925
	`Logistic Regression`	54.569
	`Bag Of Words`	53.796