Week 2
In order to firm up the original project proposal, this week was mainly spent in doing research online. The research was done in order to determine what machine learning approaches would best fit the data I’d been given, as well to get familiar with the language I’d be working with. The work that my lab does is mostly in R, so even though I had the option to choose either Python or R for the project, I ultimately decided on R in order for my project to better link up with the work the lab is currently doing.
The first thing I did was try and familiarize myself with machine learning in R, in order to better determine what potential packages or algroithms I could use. I followed through the entirety of the links labeled with the numbers 1 and 2 in the citations section below, which helped me narrow down which techniques I wanted to use. The first link focuses on implementing neural network models using the keras package in R, which I ultimately decided to use for my project due to the ability to fine tune parameters in order to get a better fitting model. In addition to this, I chose to use the k-nearest neighbors (knn) method and support vector machine (svm). I talked to other researchers within my lab, and of the methods I tried they recommended using these based on successes with other projects. These methodologies also appealed to me based on the fact that I’ve learned about them in depth in my artiificial intelligence course, which I feel gives me a better grasp of how to apply them to the given dataset.