A user freindly module with which users can just drop their dataset and download the best ML model for their dataset
Project description
Rules and guide lines for uploading the dataset:
- The file should be either .csv or .xlsx
- Number of columns : 3 < cols > 100
- Number of rows : 200 < rows > 2500
- The index col must be the first column.If the dataset doesn't have an index column include it.For example,you can use row number as index.
- The dependent variable or the target class should be the last column
Model default settings: chi square Test p val < 0.1
Train Test Validation split ratio ** 70:20:10 SSS No.of folds ** 10
Random search params scores = AUC,precision,accuracy refit criterion = AUC
KNN params: 2 < n_neighbors < 5 metric = euclidean,manhattan,minkowski
Logistic Regression: penalty = l1,none solver = default c = 0.1 geomspace,no.of elements =3
SVC params = {'C' : [1,10,100], 'kernel' : ['rbf', 'linear'], 'gamma' : ['scale', 'auto']}
Random Forest Classifier params = {'n_estimators' : [10,100,200], 'criterion' : ['gini', 'entropy']}
Decision Trees params = {'criterion' : ['gini', 'entropy'], 'splitter' : ['best', 'random']}
Naive Bayes(Gaussian) default parameters
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.