Atrial fibrillation (AF) is one of the most common arrhythmias with challenging public health implications. Therefore, automatic detection of AF episodes on ECG is one of the essential tasks in biomedical engineering. In this paper, we applied the recently introduced method of compressor-based text classification with gzip algorithm for AF detection (binary classification between heart rhythms). We investigated the normalized compression distance applied to RR-interval and $\Delta$RR-interval sequences ($\Delta$RR-interval is the difference between subsequent RR-intervals). Here, the configuration of the k-nearest neighbour classifier, an optimal window length, and the choice of data types for compression were analyzed. We achieved good classification results while learning on the full MIT-BIH Atrial Fibrillation database, close to the best specialized AF detection algorithms (avg. sensitivity = 97.1\%, avg. specificity = 91.7\%, best sensitivity of 99.8\%, best specificity of 97.6\% with fivefold cross-validation). In addition, we evaluated the classification performance under the few-shot learning setting. Our results suggest that gzip compression-based classification, originally proposed for texts, is suitable for biomedical data and quantized continuous stochastic sequences in general.

The ALS algorithm introduced by Hu et al., is a very popular technique used in Recommender System problems, especially when we have implicit datasets (for example clicks, likes etc). It can handle large volumes of data reasonably well and we can find many good implementations in various Machine Learning frameworks. Spark includes the algorithm in the MLlib component which has recently been refactored to improve the readability and the architecture of the code. Spark's implementation requires the Item and User id to be numbers within integer range (either Integer type or Long within integer range), which is reasonable as this can help speed up the operations and reduce memory consumption. One thing I noticed though while reading the code is that those id columns are being casted into Doubles and then into Integers at the beginning of the fit/predict methods.