Ezyme Classifier Using ML

Projct aims at classifying FASTA format protien sequences using multiclass classification algorithms in Spark.

Steps to final classification

1. Data Preparation through python script
2. Convert raw data csv and then to libsvm in R
3. Run spark ML algos in jupyter notebook

Spark-MLlib Algorithms Used

Logistic Regression Classifier

Decision Tree Classifier

Random Forest Classifier

As per our evaluation, this gives us best result with accuracy of 57% with 30 trees. We can further work on this to increase number of trees, max depth and other hyper parameter tunings

Feedforward Artificial Neural Network Classifier