emr4h / spam-email-and-url-detection-using-machine-learning

Detection of spam emails and spam urls by classification with machine learning

Spam-Email-and-Url-Detection-Using-Machine-Learning

About The Project

Detection of spam emails and spam urls by classification with machine learning

Getting Started

The data sets used in the project were taken from the addresses below

• Spam Email Data Set ➡️ https://www.kaggle.com/mfaisalqureshi/spam-email?select=spam.csv
• Spam Url Data Set ➡️ https://www.kaggle.com/shivamb/spam-url-prediction

Thank you to the friends who provided these datasets.

Dataset

• This dataset (spam.csv) contains a total of 5572 data. There are %83 safe and %17 spam.
• This dataset (url_spam_classification.csv) contains a total of 148.303 data. There are %68 safe and %32 spam.

Classification

Classification algorithms used for spam email detection :

* Decision Tree
* K-Nearest Neighbors
* Random Forest
* Support Vector Machine

Classification algorithms used for spam url detection :

* Decision Tree
* Stochastic Gradient Descent
* Multinomial Naive Bayes
* Linear Support Vector

Spam Email Results

The Success Rate was calculated as % : 93.89806173725772 with the K-Nearest-Neighbors
The Success Rate was calculated as % : 97.05671213208902 with Random Forest
The Success Rate was calculated as % : 96.76956209619526 with Decision Tree
The Success Rate was calculated as % : 97.98994974874373 with Support Vector Machine

Spam Url Results

The Success Rate was calculated as % : 98.80245981227749 with Decision Tree
The Success Rate was calculated as % : 98.52465206602655 with LinearSVC
The Success Rate was calculated as % : 95.27457115114899 with Stochastic Gradient Descent
The Success Rate was calculated as % : 91.10206063221491 with Multinomial Naive Bayes