Email Spam Classifier This project is an Email Spam Classifier that uses a HashMap data structure to categorize emails as spam or not spam based on word frequencies from a CSV file. It supports two hashing techniques: Chaining and Open Addressing.
Table of Contents
Project Overview
How to Use?
Technical Details
Project Overview
The Email Spam Classifier reads a CSV file with precomputed word frequencies (both spam and ham) and builds a HashMap-based structure to store this information. It then allows the user to input an email, which is analyzed based on word frequencies to classify the email as "Spam" or "Not Spam."
The classification process uses two hashing implementations:
Chaining HashMap: Uses linked lists to handle collisions. Open Addressing HashMap: Resolves collisions through linear probing.
How To Use?
Make sure that GTK is setup on your PC. Clone the git repo. Run the file main.cpp On running the code, A window will get opened where the user will be asked to enter the email he/she wants to classify. After entering the email, the user will click on the classify button./nA new window will get opened which will have the results of both the methods that we have implemented(Chaining and Open Addressing).
Technical Details:
We have Implemented 2 hash functions(Chaining and Open Addressing). The HashMap consists of the Word, its spam_count and its ham_count. For each word in the email, we calculate its spam_score (spam_score = spamFreq/(spamFreq + hamFreq)) and then take the average of spam_scores of all the words in the email. If this value is greater than the threshold, the Email is classified as Spam otherwise Ham. Also, it is possible that a new word which is not present in the HaspMap is encountered in the emial. In such cases, we assign that word, a spam score equal to the threshold value.
Logic For HashFunction: hashVal = 37 * hashVal + c. 37 is choosen because 37 is the next larger number which is greater then 26(number of alphabets) and is prime. Alternatively, we could have choosen 31 or 41 also, but while searching on internet, we found out that 37 is a better choice. Read the below article. https://medium.com/@chigwel/the-enigma-of-number-37-why-this-prime-number-is-our-intuitions-favorite-168e8947fb3e