Algorithm 1. AI-based Cooperate Meta-Model Fall Risk Prediction Algorithm: Pseudocode of the Proposed Model (Figure 4)

Input Mapping: Let: $X_{AI1}$ = Output of ${AI}_1$ Model—Normal, Low, Moderate, High, Emergency $X_{AI2}$ = Output of ${AI}_2$ Model—Low, Moderate, High Assign weights $W_{AI1}$ and $W_{AI2}$ to the outputs:                    For $X_{AI1}$: $W_{AI1}=\left\{\begin{array}{c}0.2 if Normal\\ 0.4 if Low\\ 0.6 if Moderate\\ 0.8 if High\\ 1.0 if Emergency\end{array}\right.$                    For $X_{AI2}$: $W_{AI2}=\left\{\begin{array}{c}0.4 if Low\\ 0.6 if Moderate\\ 0.8 if High\end{array}\right.$ 2.Combined Risk Score: The combined risk score $S_{combined}$ is calculated as: $S_{combined}$ = ${\displaystyle \frac{W_{AI1}+W_{AI2}}{2}}$ 3.Define Risk Levels: Risk level R is defined based on $S_{combined}$:                    $R=\left\{\begin{array}{c}Low if 0\le S_{combined}\le 0.40\\ Moderate if 0.41\le S_{combined}\le 0.80\\ High 0.81\le S_{combined}\le 1.00\end{array}\right.$ 4.Train Random Forest Meta-Model Let the dataset consist of Training features $X_{train}=\{W_{AI1}, W_{AI2}\}$ Training labels $Y_{train}$ = MFS-based risk levels {Low, Moderate, High} 5.Random Forest Model: Initialize T, the number of decision trees. n_estimators = 100 max_depth = 10 min_samples_split = 4 For each tree t in T: Select a random subset $X_t$ from $X_{train}$ Build a decision tree using $X_t$ and $Y_{train}$ Split data at each node using features $\{W_{AI1}, W_{AI2}\}$ Aggregate predictions from all T trees by majority voting. 6.Predict Fall Risk: For testing data $X_{test}=\{W_{AI1}, W_{AI2}\}$ Pass each test instance through all T decision trees. Compute the predicted risk level $R_{pred}$ based on a majority vote from the T trees. 7.Output Results For each participant i, output: Participant i: {$S_{combined}$,$R_{pred}$}