its day 1. not precisely day 1 in studying ML however extra like day 1 in cleansing the mud in my mind after I ended studying for some time.
Right now’s success job:
- Mannequin Analysis
- Hyperparameter tuning
- Day 1 undertaking — buyer churn (full pipeline)
what Mannequin Analysis Metrics?
good query. I knew some metrics from my earlier studying akin to accuracy rating, confusion matrix, classification report and r2 rating. from as we speak’s session although I added extra to the checklist akin to MAE, MSE and with cross validation.
the Hyperparameter tuning half was great point to be taught.
ended the day with buyer churn undertaking. the dataset customer churn is used. I checked the dataset for lacking values. there have been none. so i went forward, created a Pipeline to make my job straightforward as a result of i used to be gonna do Column transformation of OneHotEncoder to encode categorical columns. after that I used RandomForestClassifier to coach the mannequin. the accuracy was round 75%. so I went forward to attempt to tune the hyperparameters. I used RandomizedSearchCV for it. the param_distribution had been n_estimators, max_depth, min_samples_split. the accuracy turned 78%. there may be nonetheless room for imporvement. thats what i’m going to do for day 2.
Day 1 efficiently accomplished 🎉
