Merge pull request #6 from ShrutiC-git/lda

Lda

Author: ShrutiC-git

dvc.lock

@@ -28,16 +28,16 @@ train_data:
     md5: f4187b5e2d213d442bde7b3274cb9f68
     size: 4545
   - path: train_data.py
-    md5: 21cf343fea969cd5a3337f1c91e95a0b
-    size: 1230
+    md5: 1946a3fa319477cb4ba126b73edf0658
+    size: 1408
   params:
     params.yaml:
       cv: 5
       solver: liblinear
   outs:
   - path: final_model.sav
-    md5: 0923ea6bf844d92d94afd63eda9aff19
-    size: 329
+    md5: 9dea20b3b6ef1af445348ada4b6ba474
+    size: 200
   - path: metrics.json
-    md5: 503094384d8da9040573b1f8dc8c38bc
-    size: 88
+    md5: 5a0c41369fb24ac176368bf63fd8622e
+    size: 102

metrics.json

@@ -1 +1 @@
-{"accuracy": 0.738562091503268, "specificity": 1.0, "sensitivity": 0.012345679012345678}
+{"accuracy": 0.7581699346405228, "specificity": 0.9466666666666667, "sensitivity": 0.2345679012345679}

train_data.py

@@ -10,6 +10,8 @@
 from sklearn.preprocessing import Normalizer
 import yaml
 import pickle
+from sklearn.naive_bayes import MultinomialNB
+from sklearn.discriminant_analysis import LinearDiscriminantAnalysis
 
 df = pd.read_csv('haberman_processed.csv', index_col=0)
 
@@ -20,7 +22,9 @@
 X = Normalizer().fit_transform(X)
 
 #Using a Logistic Regression in production
-clf = LogisticRegression(solver=yaml.safe_load(open('params.yaml'))['solver'])
+#clf = LogisticRegression(solver=yaml.safe_load(open('params.yaml'))['solver'])
+#clf = MultinomialNB()
+clf = LinearDiscriminantAnalysis()
 y_pred = cross_val_predict(clf, X, y, cv = yaml.safe_load(open('params.yaml'))['cv'])
 
 # Metrics for comparing performance bw models