ECE 470 Project
- Since the SMS dataset is small, it is including in its entirely in the repo.
The motivation for this project is to develop a system that can filter out spam messages sent over SMS. The spam filter will be targeted towards unwanted marketing or phishing messages. To help select relevant features to train the model, a genetic algorithm will be used with feature extraction to find the optimal set of features used for filtering messages. Feature extraction is one of the more difficult steps in machine learning, but is essential to having accurate and reliable machine learning models. Using genetic algorithms the optimal features can be extracted from the entire feature set. Some disadvantages of genetic algorithms are computationally expensive and algorithms takes a long time to converge.
The programming languages used in this project will be C++ and Python. Libraries for parsing text, text files, and mathematical functions will be added as necessary with a preference to use standard libraries whenever possible.
[1]: Undergraduate students, Department of Electrical and Computer Engineering, University of Victoria