slugComparative Study on Email Spam Classifier Using Feature Selection Techniques
Journal Title: International Journal for Research in Applied Science and Engineering Technology (IJRASET) - Year 2014, Vol 2, Issue 6
Abstract
Now a days e-mail is very popular because the way of communication method is very easy and due to these reasons some advertisers and social networks sent messages,for advertising their product,that are unwanted for users and not requested by the users, these messages are called spam messages. Sometimes these spam messages are harmful for system after clicking the link which is sent by the spammer or advertisement companies. It takes lots of time for deleting and also occupy memory for storage. Due to these we need filtering the messages and this is the important task for separating the genuine messages from the junk messages. Even though number of researchers are uses different classification techniques for classify the spam, still 100% accuracy in spam classification are question mark, In this paper, we used spam data set which are collected from UCI repository. Initially, various classification algorithms are applied over this dataset using CLEMENTINE data mining tool. This data set is divided into two parts one is training data set and the other one is testing data set. After that most of the data is used for training and a smaller portion of the data is used for testing. After a model has been processed by using the training set, we test the model and identify the results. This process is done by different data set. Finally, best classifier for email spam i.e neural network is identified based on the Training and testing accuracy of various models. In this thesis we also use feature selection method selecting features of spam data set which removes the redundant , irrelevant and noisy data. It improves the data quality and also increase the accuracy of the resulting models.
Authors and Affiliations
Yukti Kesharwani, Shrikant Lade, Dayashakar Pandey
Keywords
Related Articles
- EP ID EP18343
- DOI -
- Views 306
- Downloads 12
Optimization of Backup Storage by Reducing Fragmentation in Distributed Environment
In modern backup systems, Deduplication plays a vital role in the elimination of duplicate data in a storage system which one of the technique to reduce storage costs. Deduplication divides a backup stream into variable...
The PolyVernam Cipher
Cryptography encryption is an effective way to achieve the security of data. The encryption is to hide the data in a way that an attacker cannot hack the data. The main purpose of encryption is to hide the data from the...
An Improved Data Transmission Protocol for VANETS Using Link State Information
Efficient data delivery in vehicular ad hoc networks (VANETs) with high mobility is a challenging issue due to dynamic topology changes and unstable wireless links. The opportunistic routing protocols can improve the re...
Path Planning in Outdoor and Unknown Environment using BAPSO Algorithm
The path planning through BAPSO algorithm in outdoor and unknown environments is a novel approach for robotic path planning. The BAPSO is a hybrid (BAT and PSO) algorithm. The main challenges of robotics are its automat...
Power management strategies of a Wind/hydrogen/ultra-capacitor based hybrid power system
The generated power does not depend on the grid requirement but entirely on the fluctuant wind condition. A dc-coupled wind/hydrogen/ultra-capacitor hybrid power system is studied in this paper. The purpose of the contr...