Sensor-Based Datasets for Human Activity Recognition – A Systematic Review of Literature

Adapted Binary Particle Swarm Optimization for Efficient Features Selection in the Case of Imbalanced Sensor Data

“…The initial values of min a and max a are equal to +∞ (line 11) and −∞ (line 12). For each feature f from the set of features F selected , the values of min a and max a are updated in (lines [14][15][16][17][18][19] as follows: If the value of M f ,a is less than the value of min a (line 14) then min a is initialized to the value of M f ,a (line 15) and if the value of M f ,a is greater than the value of max a (line 17) then max a is initialized to the value of M f ,a (line 18). If F selected = ∅ (line 21) then the value of ∆ is incremented with |max a − min a + 1| (line 22).…”

“…In the literature there are many datasets generated by sensors [15], which are used in the prediction of DLAs; however, the identification of the most appropriate techniques that return the best results, in various contexts, is still challenging due to the many dimensions that must be considered, such as the selection of the features, the balancing of the samples and the selection of the classifiers applied in the classification of the DLAs.…”

Evaluating the Impact of a Two-Stage Multivariate Data Cleansing Approach to Improve to the Performance of Machine Learning Classifiers: A Case Study in Human Activity Recognition

“…A compilation of standard time and energy domain features were extracted from the windowed data to obtain relevant information and to represent the characteristics of various activity signals. The features included the mean, natural logarithm, exponential, maximum, minimum, standard deviation, variance, root mean square (RMS), signal magnitude area (SMA), range, and median for the x, y, z axis and signal magnitude vector (SMV), and the cross correlation for each axis, as it has been suggested that these features are suitable for activity recognition [13,60]. In total, 51 features were extracted.…”

“…This was also calculated in the corresponding time window. The statistical features mentioned are common due to their simplistic nature and significant performance across a variety of activity recognition problems [13,60]. The maximum, minimum, and range features can assist in differentiating between activities that contain movements comprised of different ranges.…”

Applying Machine Learning Techniques in Older People Activity Recognition usingWearable and Mobile Devices

“…Reconhecer atividades humanas (HAR: Human Activity Recognition) e identificar (quebra de) padrões comportamentais de indivíduos é um campo de pesquisa promissor e desafiador, com aplicações em segurança, saúde, entretenimento, entre outras [8]. Apesar do crescimento da população idosa em todo o mundo, poucas pesquisas em HAR empregam datasets compostos por atividades realizadas por idosos [3,9] -o que é problemático dado que diferenças significativas foram identificadas, inclusive, entre grupos de idosos [6] Alguns pesquisadores investigam HAR a partir de dados de sensores de smartphones [6,7]. Entretanto smartwatches, dotados de acelerômetro, giroscópio, microfone e sensor de batimentos cardíacos, permitem coleta informações de forma contínua e não intrusiva [2,4].…”