/End-to-end.MusicGen-Classification

An End-to-end Project to Classify the Type Of Music 🦘

Primary LanguageJupyter NotebookApache License 2.0Apache-2.0

End-to-end.MusicGen-Classification 🎭

Streamlit App

Image

Project Contents 📦

  1. Music Data Analysis
  2. Models That Used
  3. Things weren't adding any value
  4. Why Is The Acuuracy around 60%?
  5. Deployment Step
  6. Lastly

🧐 1. Music Data Analysis

A Plots About The Class[The Target] To Get Insights

Class

⚖️ 2. Models That Used

2.1- Random Forest 👉🏻 The best algorithm

Here is an evaluation about accuracy, performance & feature importance of Random Forest

RandomForest

2.2- LGBM

Here is an evaluation about accuracy, performance & feature importance of LightGBM

LGBM

🔰 3. Things weren't adding any value

I have tried lots of models, tools, and thinking outside the box to enhance the performance and ⬆ the accuracy.

Some examples include:

  • Deeplearning Models like DatRetClassifier
  • Using Optuna to optimize the parameters
  • Implementing normalization methods
  • Employing an oversampling strategy

🤔 4. Why Is The Acuuracy around 60%?

The accuracy is still around 50-60% due to the following reasons:

  • I have dropped the categorical attribute instead of using label encoder
  • There is no correlation between data points and the class
  • Imbalanced classes, and so on

🧑‍🚀 5. Deployment Step

A snippet pictures about the app 👇🏻

deployment

deployment

deployment

deployment

🌜 6. Lastly

Give a try to my app 👉🏻 Streamlit App

Watch the app life on my youtube channel 👉🏻 https://youtu.be/LU80ixSVQ-c?si=koPdALvQjMxaznpg