Fake Ad Detection Using Text Classification

Abstract

In this Project we tend to explore the concept of text classification and data scrapping to detect the fake ads posted on Pakwheels.com, basically two types of ads which a normal person is unaware of, fake and genuine ads, fake ads are the ones posted by dealers and genuine ads are posted by original sellers, by using scrapping algorithms and the concept of text classification we will be detecting the fake ads. More the 90% people are dealers on Pakwheels.com. The system will apply the series of algorithms in achieving this goal. Data preprocessing, feature extraction, training a model. We will be training the model using the different models, Naïve Bayes, Support Vector Machine (SVM), Neural Network, Long Short-Term Memory (LSTM) and the algorithm with the best accuracy will be used in our system.