-
Motivation: Build my first project in the data science space. After having read a bit of theory and some books about Machine Learning and Convolutional Neural Netowrks, and got to know about the UCI ML Repository, I set out to apply my knowledge to some dataset and make a project out of it. I came across the Abalone dataset and read the background behind this classification problem to solve it.
-
Challenge: Getting a good enough accuracy in the built ML model. The problem with the dataset is that it is small. Nevertheless, attempts have been made to eke out every possible increase in accuracy using different classification models.
-
Accomplishment: Approaching the problem through a different approach and publishing a research paper documenting this approach and the results obtained using the same. Given that the target values in the dataset are numeric with a sense of ranking, I could in theory try regression techniques on this classification problem. After doing EDA, that is what I did and then published a paper on it in the IJCA.
- Python 3
- pandas
- matplotlib
- numpy
- scikit-learn
You can view the IJCA version of the published paper here