Table 2 Scenario based evaluation of ADDT.

From: Enhancing wearable sensor data analysis for patient health monitoring using allied data disparity technique and multi instance ensemble perceptron learning

Scenario

Clinical setting

Challenge

Baseline method

(ADDT + ensemble perceptron)

Improvement/observation

S1: real-time ICU monitoring

Intensive Care Unit with fluctuating vitals (HR, BP, SpO₂)

Sensor dropout due to noise/interference

Rule-based alerts with threshold-only decision

Identifies substitute clinical data using disparity detection and maintains continuity via perceptron fusion

Reduces false alerts by 35%, improves sequence consistency by ~ 28%

S2: remote patient monitoring

At-home care for elderly patients with wearable ECG & oxygen sensors

Data gaps due to poor connectivity or motion artifacts

Traditional interpolation or missing-value imputation

Disparity detection flags unreliable data; ensemble adapts using highest-confidence clinical matches

Increases prediction accuracy by ~ 21%, lowers anomaly detection time by ~ 18%

S3: chronic disease monitoring

Diabetic patient with wearable glucose and pressure sensors + periodic clinical reports

Time-misaligned sensor-clinical inputs

Linear regression or static ML models

Periodic correlation adjusts weights with clinical trends; adaptive precision-based update

Achieves 12–16% better glucose level classification; improved adaptability to patient variation

Back to article page