Detect input drift with machine learning algorithms with python
Budget: $250 – $750 CAD
Our Model Behavior: Supervised ML with monthly batch prediction (batch of clients). The model provides a score of the customer’s behavior related relating to their credit card usage. This score lets us know if the client is a good or bad for limit increase / decreasing purposes.
I need to monitor the input data of Model Behavior.
The input data that was used to build and train the model could change with the time.Example :the distribution of the feature age could change with the time.
I need to detect if the distributions of the feartures of the production are different of the distributions of the features of the baseline.
To detect the drift, I need to use :
1-SVM, Random Forests and KNN
2-CNN : capture the distributions as images and then compare the distributions to say if there is drift or no.
Please see the word document for more explication
I need to monitor the input data of Model Behavior.
The input data that was used to build and train the model could change with the time.Example :the distribution of the feature age could change with the time.
I need to detect if the distributions of the feartures of the production are different of the distributions of the features of the baseline.
To detect the drift, I need to use :
1-SVM, Random Forests and KNN
2-CNN : capture the distributions as images and then compare the distributions to say if there is drift or no.
Please see the word document for more explication