Introduction

This is the code implementation for Spectra classification. We use machine learning methods to predict whether an unseen sample contains a particular or several species. The code includes various machine learning methods which are involved in our experiments, including Neural Networks, Support Vector Machine, Naive Bayes, K-Nearest Neighbors and more. The experiment results are reported in Experiment results.

Data

The data distribution that we used in our experiments is shown as below.

Experiment results

Here are some of our experiments results.
Our model performence on detecting whether a sample contains a certain species is shown as below.

Here are the comparisons of two versions of our models under different experiment settings.

Contributors

Chathurika Rathnayake
Zhenghao Zhao
Yan Yan
Rong Wang