Human Gait Analysis: A Sequential Framework of Lightweight Deep Learning and Improved Moth-Flame Optimization Algorithm

Table 4

Confidence interval-based analysis of proposed framework on the TUM GAID dataset.

Confidence level	Margin of error
Normal walk
68.3%,	98.767 ± 0.62 (±0.63%)
90%, 1.645	98.767 ± 1.02 (±1.03%)
95%, 1.960	98.767 ± 1.215 (±1.23%)
99%, 2.576	98.767 ± 1.597 (±1.62%)
Walk with a bag
68.3%,	97.71 ± 0.573 (±0.59%)
90%, 1.645	97.71 ± 0.942 (±0.96%)
95%, 1.960	97.71 ± 1.123 (±1.15%)
99%, 2.576	97.71 ± 1.475 (±1.51%)
Walk with a coat
68.3%,	96.925 ± 0.513 (±0.53%)
90%, 1.645	96.925 ± 0.843 (±0.87%)
95%, 1.960	96.925 ± 1.005 (±1.04%)
99%, 2.576	96.925 ± 1.321 (±1.36%)