*2.3. Algorithms*

There are two categories of algorithms: *threshold-based* and *ML-based*. Threshold algorithms simply define limit values, outside of which, a fall is detected. They have often been sufficient but they tend to produce false alarms especially with fall-like activities such as sitting abruptly [16]. To compensate, these studies [13,22] added simple posture and pattern recognition algorithms that detect changes in body posture and level of activity. This improves the detection's robustness while keeping a low computational complexity. However, it may still fail during specific falls and ADLs. For example, Sucerquia et al. [31] used a manual threshold-based classification over their dataset *SisFall*, achieving 96% accuracy.

ML algorithms automatically learn patterns based on data, and very commonly include feature extraction. They require more computational power and are complex to optimize but produce improved results. Most of the studies such as [11,13] employed a supervised learning technique. Common algorithms are k-Nearest Neighbor [27], Support Vector Machine [18,27] and Artificial Neural Network [11,27]. Yuwono et al. [15] used unsupervised learning which works with clusters. This is a compelling solution because it does not require labeled data. The state-of-the-art Deep Learning algorithms are increasing in popularity, achieving promising results in various fields. Musci et al. [36] employed Recurrent Neural Networks to detect falls. They used a publicly available dataset (*SisFall*) [37] and reported outperforming the results of the original paper [31]. Casilari et al. [35] employed a Convolutional Neural Networks on several datasets, including *SisFall* [31]. They reported promising results with a Sensitivity and Specificity over 98%.
