You will submit your final code online, as a single zip file. The zip file should include:– a short presentation (no more than 8 slides) that describes what you did.– a main script (called runner.py) that implements your best classification scheme (i.e. the schemethat gave you the best accuracy). – one or more scripts (this is up to you) that implement other things (algorithms, features, parameters) that you tried. These are important, even if they didn’t achieve the best possible accuracy.
You will be evaluated based on both your best accuracy and the number/nature of the various classificationschemes that you tried. I suggest you experiment with as many many algorithms and parameter configurations as possible.
Your scripts will be trained and tested on two files that I create myself for the purposes of the project. You can assume that both the training and testing files will be subsets of the entire dataset that you are working on. The files will have the following format:– one article per line – each line will have 2 tab-separated columns. The first will be the article’s text (we won’t be considering any other features included in the dataset). The second will be the label (the publication).– You can assume that my files will not include more than 5 distinct labels.