Algorithm 1 Prediction algorithm

Input: AccelerometerData ($A_x,A_y,A_z$) in CSV file Output: prediction of true or false of a fall $slidingWindow=750$ ms $consecutiveCount=0$ $prediction=false$ for all$AccelerometerData$ in $slidingWindow$ do   Compute $A_{\mathrm{res}}$, $\Delta S$, $S_{min}$, $S_{max}$   Write to CSV file end for Initialize the prediction interval (i.e., which data sample in the CVS file to start the prediction) traditionModel = the trained NB or SVM fall detection model for all$instance$ in $predictionInteval$ do   $label$ = traditionalModel.classifyInstance($instance$)   if ($label==″Fall″$) then     ++$consecutiveCount$   else if ($3<=consecutiveCount<=50$) then     $prediction=true$     $consecutiveCount=0$   else     $consecutiveCount=0$   end if end for return$prediction$