Converting the logistic regression Scikit-learn model trained to the Core ML Format.

Author: DevOpsThinh

usecases/tabular_classifier/logistic_regression/converting_scikit-learn_model_coreml.py

@@ -0,0 +1,32 @@
+# Author/ Learner: Nguyen Truong Thinh
+# Contact me: [email protected] || +84393280504
+#
+# Use case: Create a Logistic Regression model that can be used to convert into
+# the Core ML Format via CoreML Tool .
+# The model will be trained & converted on the popular UCI ML Pima Indians Diabetes dataset.
+
+import coremltools
+
+from usecases.tabular_classifier.logistic_regression.logistic_regression_classifier import trained_model
+
+coreml_model = coremltools.converters.sklearn.convert(trained_model, ['Pregnancies', 'Glucose',
+                                                                      'BloodPressure', 'SkinThickness', 'Insulin',
+                                                                      'BMI', 'DiabetesPedigreeFunction', 'Age'],
+                                                      'Outcome')
+
+coreml_model.author = 'Nguyen Truong Thinh'
+coreml_model.short_description = 'A logistic regression model trained on the Kaggle.com version of th Pima Indians ' \
+                                 'diabetes dataset.'
+# Features description
+coreml_model.input_description['Pregnancies'] = 'Number of pregnancies.'
+coreml_model.input_description['Glucose'] = 'Plasma glucose concentration after 2 hours in an oral glucose tolerance ' \
+                                            'test.'
+coreml_model.input_description['BloodPressure'] = 'Diastolic blood pressure.'
+coreml_model.input_description['SkinThickness'] = 'Thickness of the triceps skin folds.'
+coreml_model.input_description['BMI'] = 'Body mass index.'
+coreml_model.input_description[
+    'DiabetesPedigreeFunction'] = 'A function that determines the risk of diabetes based on family history.'
+coreml_model.input_description['Age'] = 'The age of the subject.'
+# Description of target variable
+coreml_model.output_description['Outcome'] = 'A binary value, 1 indicates the patient has type-2 diabetes.'
+coreml_model.save('diabetes_indian.mlpackage')

usecases/tabular_classifier/logistic_regression/diabetes_indian.mlpackage/Data/com.apple.CoreML/model.mlmodel

@@ -0,0 +1,12 @@
+{
+    "fileFormatVersion": "1.0.0",
+    "itemInfoEntries": {
+        "F2D67E27-AF2E-4D54-89B0-9990C4DF7371": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Specification",
+            "name": "model.mlmodel",
+            "path": "com.apple.CoreML/model.mlmodel"
+        }
+    },
+    "rootModelIdentifier": "F2D67E27-AF2E-4D54-89B0-9990C4DF7371"
+}

usecases/tabular_classifier/logistic_regression/diabetes_indian.mlpackage/Manifest.json

@@ -48,7 +48,7 @@
 
     # Train a logistic regression model
     model = LogisticRegression(penalty='l2', fit_intercept=True, solver="liblinear", multi_class='ovr')
-    trained_model = model.fit(df_diabetes_features_train, df_diabetes_target_train.values.ravel())
+    trained_model_kfcv = model.fit(df_diabetes_features_train, df_diabetes_target_train.values.ravel())
 
     # Get predictions
     predictions = model.predict(df_diabetes_features_test)