Multi Stage Neural Network Based Ensemble Learning Approach for Wheat Leaf Disease Classification

Exponential growth of mobile advertising, fraudulent ad clicks have become a major concern, leading to substantial financial losses for advertisers. This project, Ad Click Fraud Detection Using Machine Learning and Deep Learning Algorithms, aims to develop an intelligent and scalable detection system capable of distinguishing between genuine and fraudulent clicks. The dataset contains crucial features such as IP address, application ID, device type, operating system version, publisher channel ID, click timestamp, and attribution time (if the app was downloaded post-click). The target variable is attributed indicates whether a click led to an actual app download. A wide range of models were implemented, including traditional Machine Learning algorithms like Logistic Regression, Decision Tree, Random Forest, K-Nearest Neighbors, Naive Bayes, SVM, Gradient Boosting, LightGBM, and XGBoost. Additionally, Deep Learning architectures such as Artificial Neural Networks (ANN), Convolutional Neural Networks (CNN), Deep Neural Networks (DNN), Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU) were explored. A Stacking Classifier was used to combine model strengths and optimize performance. The backend integrates all models to compare results, while the Flask-based frontend allows users to input click features and receive real-time fraud detection outputs.

Keywords: Ad Click Fraud, Machine Learning, Deep Learning, Flask App, Stacking Classifier, LSTM, XGBoost, Click Prediction, Mobile Advertising, Classification.