ABALONE PROJECT

This Repository Abalone classification project. This is the first project in my self learning curriculum.

Dataset

The dataset I have used can be found here (https://www.kaggle.com/rodolfomendes/abalone-dataset)

Purpose of project

In the journey of learning machine learning especially classification EDA and classification model evaluation, I have made a self learning curriculum which is comprised of building 10 classification projects end-to-end involving steps, which are as follows:

• Do the EDA

  • Using pandas, numpy, statsmoel, sklearn, seaborn and matplotlib
  • Check for bias-sampling such as data imbalance
  • descriptive statstics
    • Measure of central tendency
    • Measure of dispersion
    • Measure of association
    • Check for skewness and kurtosis
    • If needed data imputation
    • If needed data transformation
    • Outlier detection and handling

• Choose best Model

  • Train model systematically
  • Use ensemble models
  • Use cross-validation method to reduce variance error

• Do the model evaluation

  • metrics: accuracy, precision, recall, F1-score and ROC-AUC
  • use mlxtent to observe bias variance decomposition of error
  • AIC and BIC for checking model bias-variance

Building model API

Using flask library build model-api. Learn the folder structure of api backend which can be scaled later if required.

Deploying API

Learn deplyment to digitalocean and deploying to heroku from github.