Online semi-supervised multi-channel time series classifier based on growing neural gas

Abstract

Challenges in time series classification has attracted attention in the past decade. Although large amounts of labeled data are assumed to be available, in reality, labeled data might be scarce to find in many domains. In this paper, we propose an online semi-supervised multi-channel classifier for time series based on growing neural gas (GNG) learning scheme. The method is able to handle multi-channel time series with variation in dimensions and it introduces a label prediction strategy to minimize misclassification. It measures the similarity of input instance and learned templates using weighted multichannel dynamic time warping technique and learns the topology of input data space specified for each class using the GNG learning algorithm. Comprehensive evaluation is conducted using various datasets, such as gesture recognition, human activity recognition, and human daily-life activity recognition. Experimental results demonstrate good classification results, with indication that the proposed approach requires only a handful of labeled instances to construct an accurate classification model.

Improving robustness against electrode shift of sEMG based hand gesture recognition using online semi-supervised learning

“…Due to this flexibility, online semi-supervised learning has been applied to many applications, e.g. human activity recognition and human daily-life activity recognition [13], face recognition [16] and tracking applications [17], with satisfying performance.…”

“…OSSMGNG proposed by Parham Nooralishahi et al [13] is an incremental learning method based on growing neural gas. This method improves the performance of a classifier by using a small amount of labeled data and a large amount of unlabeled data.…”

“…A classifier is trained with a small amount of labeled samples and then is updated with unlabeled samples online in online semi-supervised learning. Online multi-channel semi-supervised growing neural gas (OSSMGNG) [13], which is one of well-known methods, is used in this paper. OSSMGNG contains two layers: the first layer is the category layer, and the other layer is the topological layer of associated classification.…”

Time-Series Prediction of Iron and Silicon Content in Aluminium Electrolysis Based on Machine Learning

“…A Kmean dynamic clustering is proposed by Liang [25], which overcomes the hysteresis and low accuracy of conventional overflow monitoring methods. Nooralishahi [26] develops an online semi-supervised multi-channel classifier based on GNG learning scheme, to really complete machine learning online.…”