A Study of Accelerometer and Gyroscope Measurements in Physical Life-Log Activities Detection Systems

. 2020 Nov 21;20(22):6670. doi: 10.3390/s20226670

Algorithm 2: Genetic algorithm-based reweighted feature selection

Input: FV: Multi-fused feature vectors (u₁, u₂, u₃,… u_n) /* acquire feature vector */
Output: FL: Multi-fused feature list (l₁, l₂, l_m) /* obtain vector of optimal features*/
/* feature vectors are converted into corresponding chromosomes */
for vector in population_lab do
/* multi-fused-feature vectors are further processed and reweighted to extract an optimal weight
RewightedFeatures <- []
while fitness not achieved or fitness not changing do
for feature in vector do
ReweightedFeatures (feature)
end for
Rechoose ()
offspring1, offspring2 <- CrossOver (vector) /* calculate crossover chromosomes*/
mutated <- Mutation (vector) /* calculate global maxima */
/*obtain relevant features on the basis of Linear Support Vector Machine (LSVM) and random forest-based fitness function*/
Evaluationfunction <- GetFitness (vector)
end while
return ReweightedFeatures
end for