Queried the data using MongoDB, trained and validated the model using LinearRegression. Performed EDA. Also, used Feature Engineering process to increase Performance and applied PCA to get better intuition of Dimensionalty Reduction Algorithms.
Challenge OBJECTIVES:-
- Load the data file provided in a local instance of MongoDB
- Extract all data from the ‘sensor_data’ collection to form the data for ML task.
- Create a regression-based ML/DL model, to predict the 'Pump Radial Bearing Vibration' column from the data extracted.
- Share your understanding of underlying data using descriptive analytics (You’re encouraged to do data cleaning, EDA et al.)
- Use of proper feature selection process and hyperparameter tuning (if required) is expected.
- You’re allowed to do any feature engineering/feature selection you deem necessary
- Properly comment on the code as and where required
- Create a script which will load the model and make predictions for data provided to it for testing purposes (sample script attached)
Several different Model Selection, Feature Engineering to increase model score can be done.
A Beginner Friendly project if never used MongoDB for data extraction