Novel Stacking Classification and Prediction Algorithm Based Ambient Assisted Living for Elderly

Algorithm 3

Novel Stacking Classification and Prediction algorithm (NSCP)

Input: Fall Detection Training Dataset Cluster 1 (FTC1), Fall Detection

    Training Dataset Cluster 2 (FTC2), Fall Detection Testing Dataset

    (FDTD)

Output: Fall Detection Predicted Result (FDPR)

Step 1: Classify FTC1 based on RIPPER classifier using weka

Step 2: Classify FTC1 based on MLR classifier using weka

Step 3: Classify FTC2 based on Dl4jMlpClassifier classifier using weka

Step 4: For each data instance DI from FDTD

Step 5:  P1 = Predict DI using RIPPER classifier

Step 6:  P2 = Predict DI using MLR classifier

Step 7:  P3 = Predict DI using Dl4jMlpClassifier classifier

Step 8:  FDPR = Predict DI using Stacking classifier (RIPPER + MLR +

    Dl4jMlpClassifier as base classifier and Naïve Bayes as Meta-

    Classifier)

Step 9: End For