Apply ANN for learning a set of features of phishing email, and further classify the NEW similar attacks. It must be ONLINE that is incoming data, improves the o/p, adds to training set.
Relevant Skills and Experience
Recent works: 2 MS projects, on text based email filtering using non-recurrent neural netwroks (MLP with 2 hidden layers) for spam/ham filtering, biometric signatures based authentication with ANN
Proposed Milestones
$100 USD - First: Initial draft, confusion matrices, accuracy, ROC curves in MATLAB
$100 USD - Final: Code, revision based on feedback
Additional Services Offered
$100 USD - Documentation, assistance in running and simulating results, tutorials
Queries:
Do you have any links, additional information to share? e.g. dataset, breadth and depth of research?
Regarding project:
If the dataset is preprocessed for NLP, i.e. articles removed, in .txt or .csv format, labelled then the NN architecture can be started instantly on MATLAB or Python, as per your preference. It will be based on available set of features only. Any effort for feature engineering, is additional RnD and should have extra budget.
About me:
I'm a research student in machine learning and statistical applications to images, stock data, point clouds, time series and consult start-ups on various applications, algorthms development. I also supply students with code for Kaggle competitions.
Kindly message me whenever you're free and we can proceed. Thanks!