Transfer learning and information retrieval applied to fall detection

Abstract

Detecting falls in the elderly population is a very important issue that is related with the time of recovery. This study focuses on using wearable smart watches to monitor the movements of the user in order to detect patterns that might be related to fall events. The proposed solution explores Symbolic Aggregate approXimation (SAX) Time Series representation, together with two information retrieval techniques enriched with transfer learning (TL). The solution is user centred; that is, a model is developed for each specific user. Basically, the fall detection approach makes use of a finite-state machine to detect peaks; the time series window embedding these peaks are represented using SAX. Assuming the data from the public fall detection data sets are valid, a dictionary is prepared using the most relevant words. This dictionary is then introduced as previous knowledge to an online learning classifier that is trained with normal activities of daily living. The two classifiers are evaluated and compared with two classical approaches. Before this comparison, two clustering approaches are studied to produce the bag of relevant words. A complete experimentation is included, which makes use of several publicly available data sets and also with a data set developed by the research group. Comparisons are performed for all the data sets, showing how the TL stage empowers the classifier. The results show that this solution produces high detection rates and at the same time performed similarly for all the individuals tested. Furthermore, the positive effects of TL in this context are clearly remarked. K E Y W O R D S fall detection, machine learning, time series, transfer learning 1 | INTRODUCTION Fall detection (FD) is still a challenge concerning to the elderly population (Willy & Osterberg, 2014). The performance of the FD devices, in their different nature and applications, shows a relatively high number of false alarms or, even worse, unidentified fall events. As a consequence, these alarming devices are considered annoying, costly, and worthless.

Transfer Learning With Time Series Data: A Systematic Mapping Study

“…Symbolic aggregation approximation: Instance selection can also be carried out by selecting representations of data instances. Two of the identified publications use the symbolic aggregation approximation (SAX) representation for time series data translating each time series into a word [157], [158]. Such word representations can be collected into a bag of words and in this way form a subset of the input data.…”

“…The transfer is conducted by collecting a bag of common words, where commonness is measured by the relative term frequency. Fanez et al [158] conduct a clustering of words and only select the cluster centroids into the bag of transferred words.…”

Computer‐aided diagnosis ofCOVID‐19 disease from chest X‐ray images integrating deep feature extraction

“…Analysis of Variance (ANOVA) is an analysis tool and a statistical method used for comparison of numerous means across different groups (Fañez et al, 2020). This technique is used to test the influence of the independent variable on the dependent variable.…”

“…KNN classification is one of the most essential and basic classification strategies and ought to be one of the main decisions for a classification study when there is practically no earlier knowledge about the data distribution (Fañez et al, 2020).…”

Special issue on new trends and challenges of bio‐inspired computational intelligence algorithms in massively complex systems

“…In this case, artificial bee colony is the algorithm selected to optimize the performance of the energy management system. Detecting falls in the elderly population is a very important issue that is related with the time of recovery. This problem is tackled by Fañez, Villar, de la Cal, Sedano, & González (2020), where the proposed solution explores symbolic aggregate approximation (SAX) time series representation, together with two information retrieval techniques enriched with transfer learning. The performance of the proposed algorithm is evaluated using several publicly available data sets and also with a data set developed by the authors' research group.…”