Tensorflow Models

This repository contains two tensorflow models namely Stacked Autoencoders and Deep Belief Network.

There are various variants of both networks.

Usage

The links to the dataset are here.

Download all the files

KDD_Test

KDD_Train

KDD_Valid

KDD_TestLabels

KDD_TrainLabels

KDD_ValidLabels

Clone this repository and open the desired algorithm in any text editor to access settings

Change the path of the files

df = pd.read_csv('/home/jay/Documents/Project/incomplete_project/DBNKDD/dataset/NSL-KDD_Processed/Kdd_Test_41.csv')             # test set 
er = pd.read_csv('/home/jay/Documents/Project/incomplete_project/DBNKDD/dataset/NSL-KDD_Processed/NSL_TestLabels_mat5.csv')     # test labels
ad = pd.read_csv('/home/jay/Documents/Project/incomplete_project/DBNKDD/dataset/NSL-KDD_Processed/Kdd_Train_41.csv')            # train set 
qw = pd.read_csv('/home/jay/Documents/Project/incomplete_project/DBNKDD/dataset/NSL-KDD_Processed/NSL_TrainLabels_mat5.csv')    # train labels
tr = pd.read_csv('/home/jay/Documents/Project/incomplete_project/DBNKDD/dataset/NSL-KDD_Processed/Kdd_Valid_41.csv')            # valid set
yu = pd.read_csv('/home/jay/Documents/Project/incomplete_project/DBNKDD/dataset/NSL-KDD_Processed/NSL_ValidLabels_int3.csv')    # valid labels
rt = pd.read_csv('/home/jay/Documents/Project/incomplete_project/DBNKDD/dataset/NSL-KDD_Processed/NSL_TrainLabels_int.csv')
t = pd.read_csv('/home/jay/Documents/Project/incomplete_project/DBNKDD/dataset/NSL-KDD_Processed/NSL_TestLabels_int.csv')

You can also hard code the parameters as well as you can input them when you run the program

pre_learning_rate = float(input("Please input the Pretraining learning rate(should be between 0 and 1) : ")) 
pre_training_epochs = int(input("Please input the Pretraining epochs(more >> better) : "))
pre_batch_size = int(input("Please input the Pretraining batch size(lower >> better) : "))
display_step = 1

Also you can change the network architecture

pre_n_hidden_1 = int(input("\nPlease input the Pretraing network's Hidden layer 1'st Neurons : ")) # 1st layer num features
pre_n_hidden_2 = int(input("Please input the Pretraing network's Hidden layer 2'nd Neurons : "))# 2nd layer num features 
pre_n_hidden_3 = int(input("Please input the Pretraing network's Hidden layer 3'rd Neurons : "))
pre_n_hidden_4 = int(input("Please input the Pretraing network's Hidden layer 4'th Neurons : "))
pre_n_input = 41