ML - Algorithm (Decision Tree Classifier)

ML - Algorithm for the Analyzer Project

Supervised Algorithm for network analysis

Code Flow :

• [Step 1] - trains the algorithm (organizing data , creating grid and taking best module to use the classification process on)

• [Step 2] - organizes the test set (different csv file)

• [Step 3] - running the module on the test file

• [Step 4] - creating CSV file called Results.csv that describes the results

CSV Files usage explanation :

CSV Files	Meaning
SmallTrain	training the data on a 3MB file
BigTrain	training the data on a 14MB file
SmallTest	testing the data on a 2MB file
Results	The outcome of the algorithm with the [Class , Protocol]

Field Values for Protocol

Field	ID
ICMP	0
TCP	1
UDP	2

Field Values for Class

Field	ID
Anomaly	0
Normal	1

Todos

• Check about the SelectKbest to select best feature usage --> SelectKBest On Decision Tree .
• Check about how to present the decision tree with matplotlib / graphviz .
• Output which of the features are the one casuing it to be labeled as Anomaly.