Emotion Monitoring | Real-Time Dashboard

Player: —

ID: — · Session: 00:00:00

Game: Mobile Legends: Bang Bang

Match Type: Ranked ·

Status: Connected

Signals: 0

Battery Level: 87%

Wearable prototype (sim)

Current Emotion

😐

—

Confidence: —

Real-Time Physiological Signals

View All

Heart Rate (BPM) —

SDNN (ms) —

RMSSD (ms) —

Emotion Trend (Last 5 Minutes)

Live Monitor

Combined live view of emotion classification and vitals for the active session.

Emotion—

HR—

SDNN—

RMSSD—

LF/HF—

Physiological Signals

Performance Analytics

Match metrics

Emotion–performance summary

Emotion	n	Avg KDA	Avg gold	Win %

Model Selection

Thesis comparison: Random Forest (87.3%), ANN (85.7%), SVM (83.1%). Select a model to use for emotion classification.

Thesis highlights

Model	Accuracy	Role
Random Forest	87.3%	Primary / best overall
Artificial Neural Network (ANN)	85.7%	Deep learning baseline
Support Vector Machine (SVM)	83.1%	RBF kernel baseline

Per-emotion (RF): Focused F1 = 0.91 (highest); Frustrated F1 = 0.78 (lowest, overlap with Stressed). HRV and GSR features contributed most to classification performance.

Session History

ID	Player	Status	Started

Alerts

No critical alerts. Stress / GSR spikes will appear here during live sessions.

Settings

Prototype settings. Sensor simulator and RF model paths are documented in docs/WIRING.md.

API base: api/
Poll interval: 2s
Emotions: focused, stressed, excited, relaxed (Calm), frustrated
Dataset: ecg_hrv_emotion_dataset.csv (1,100 HRV windows · 5 emotions)
Simulator: python simulator/run_session.py