In this project we have designed and developed a Deep Neural Network with 18-layers using the DenseNet Architecture that predicts ARDS disease using Ventilator Waveform Data.
The dataset consists of raw breath data of 100 ICU patients. The model achieved an accuracy of 81% and an AUC score of 0.95.
Language and Libraries used: Python 2.7, PyTorch, Pandas, NumPy, Matplotlib
In this project we have developed and trained a Deep Neural Network with 3 hidden layers that classified a URL as Phishing or Benign using just its lexical features. The model used the help of Recurrent Weighted Average algorithm, which is an extention of the Long Short Term Memory (LSTM) RNN model. The model treats each URL as a serialized input and learn the patterns of the occurence of characters in the URL.
The dataset consists of 50,000 URLs where half are Phishing and the other half is benign.
The model achieved an accuracy of 98.6%.
Language and Libraries used: Python 3.4, TensorFlow, Pandas, NumPy, Tkinter
This projects helps the computer learn to play Atari using Q-Learning algorithm (Reinforcement Learning).
The data used is screenshots of the Atari game which is obtained using the gym library.
The trained model was successful in achieving the highest reward of 21 points in session.
Language and Libraries used: Python 3.4, PyTorch, Gym, Pandas, NumPy, Matplotlib