- Music Data Analysis
- Models That Used
- Things weren't adding any value
- Why Is The Acuuracy around 60%?
- Deployment Step
- Lastly
A Plots About The Class[The Target] To Get Insights
2.1- Random Forest 👉🏻 The best algorithm
Here is an evaluation about accuracy, performance & feature importance of Random Forest
2.2- LGBM
Here is an evaluation about accuracy, performance & feature importance of LightGBM
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
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
A snippet pictures about the app 👇🏻
Watch the app life on my youtube channel
👉🏻 https://youtu.be/LU80ixSVQ-c?si=koPdALvQjMxaznpg