Jakub Kuzilek

Electrocardiogram beat detection enhancement using Independent Component Analysis

Beat detection is a basic and fundamental step in electrocardiogram (ECG) processing. In many ECG applications strong artifacts from biological or technical sources could appear and cause distortion of ECG signals. Beat detection algorithm desired property is to avoid these distortions and detect beats in any situation. Our developed method is an extension of Christovs beat detection algorithm, which detects beat using combined adaptive threshold on transformed ECG signal (complex lead). Our offline extension adds estimation of independent components of measured signal into the transformation of ECG creating a signal called complex component, which enhances ECG activity and enables beat detection in presence of strong noises. This makes the beat detection algorithm much more robust in cases of unpredictable noise appearances, typical for holter ECGs and telemedicine applications of ECG. We compared our algorithm with the performance of our implementation of the Christovs and Hamiltons beat detection algorithm.

Independent Component Analysis and Decision Trees for ECG Holter Recording De-Noising

We have developed a method focusing on ECG signal de-noising using Independent component analysis (ICA). This approach combines JADE source separation and binary decision tree for identification and subsequent ECG noise removal. In order to to test the efficiency of this method comparison to standard filtering a wavelet- based de-noising method was used. Freely data available at Physionet medical data storage were evaluated. Evaluation criteria was root mean square error (RMSE) between original ECG and filtered data contaminated with artificial noise. Proposed algorithm achieved comparable result in terms of standard noises (power line interference, base line wander, EMG), but noticeably significantly better results were achieved when uncommon noise (electrode cable movement artefact) were compared.

Open University Learning Analytics dataset

Learning Analytics focuses on the collection and analysis of learners’ data to improve their learning experience by providing informed guidance and to optimise learning materials. To support the research in this area we have developed a dataset, containing data from courses presented at the Open University (OU). What makes the dataset unique is the fact that it contains demographic data together with aggregated clickstream data of students’ interactions in the Virtual Learning Environment (VLE). This enables the analysis of student behaviour, represented by their actions. The dataset contains the information about 22 courses, 32,593 students, their assessment results, and logs of their interactions with the VLE represented by daily summaries of student clicks (10,655,280 entries). The dataset is freely available at https://analyse.kmi.open.ac.uk/open_dataset under a CC-BY 4.0 license.

Data literacy for learning analytics

This workshop explores how data literacy impacts on learning analytics both for practitioners and for end users. The term data literacy is used to broadly describe the set of abilities around the use of data as part of everyday thinking and reasoning for solving real-world problems. It is a skill required both by learning analytics practitioners to derive actionable insights from data and by the intended end users, such that it affects their ability to accurately interpret and critique presented analysis of data. The latter is particularly important, since learning analytics outcomes can be targeted at a wide range of end users, some of whom will be young students and many of whom are not data specialists. Whilst data literacy is rarely an end goal of learning analytics projects, this workshop aims to find where issues related to data literacy have impacted on project outcomes and where important insights have been gained. This workshop will further encourage the sharing of knowledge and experience through practical activities with datasets and visualisations. This workshop aims to highlight the need for a greater understanding of data literacy as a field of study, especially with regard to communicating around large, complex, data sets.

An automatic method for holter ECG denoising using ICA

Our work aims at processing of holter ECG recordings using Independent Component Analysis (ICA). This powerful tool enables us to create a adapting and robust method for detection and estimation of noise in records. Our method is fully automatic and it is easily modifiable for any type of noise. It also preserves enough information for estimation of ECG beat type, which is a desirable feature.

Comparison of JADE and Canonical Correlation Analysis for ECG de-noising

This paper explores differences between two methods for Blind Source Separation within frame of ECG de-noising. First method is Joint Approximate Diagonalization of Eigenmatrices, which is based on estimation of fourth order cross-cummulant tensor and its diagonalization. Second one is the statistical method known as Canonical Correlation Analysis, which is based on estimation of correlation matrices between two multidimensional variables. Both methods were used within method, which combines the Blind Source Separation algorithm with decision tree. The evaluation was made on large database of 382 long-term ECG signals and the results were examined. Biggest difference was found in results of 50 Hz power line interference where the CCA algorithm completely failed. Thus main power of CCA lies in estimation of unstructured noise within ECG. JADE algorithm has larger computational complexity thus the CCA perfomend faster when estimating the components.

Processing Holter ECG signal corrupted with noise: Using ICA for QRS complex detection

Holter recording of electrocardiographic signal (ECG) is usually disturbed by noise added to measured useful signal due to e. g. worse contact skin-electrode, body movements, etc. Our general goal is to find at least some piece of information from ECG (in our case — positions of QRS complexes) in corrupted signal, on which common preprocessing methods fail. The main purpose of the article is description of an automatic method of data preprocessing for QRS detection, which is the first step in our way to final solution. We use Independent Component Analysis based ”filter” that automatically chooses noise-free components. Selection is based on morfology of noise-free ECG components.

Student Drop-out Modelling Using Virtual Learning Environment Behaviour Data.

With the rapid advancement of Virtual Learning Environments (VLE) in higher education, the amount of available student data grows. Universities collect the information about students, their demographics, their study results and their behaviour in the online environment. By applying modelling and predictive analysis methods it is possible to predict student outcome or detect bottlenecks in course design. Our work aims at statistical simulation of student behaviour in the VLE in order to identify behavioural patterns leading to drop-out or passive withdrawal i.e. the state when a student is not studying, but he has not actively withdrawn from studies. For that purpose, the method called Markov chain modelling has been used. Recorded student activities in VLE (VLE logs) has been used for constructing of probabilistic representation that students will perform some activity in the next week based on their activities in the current week. The result is an instance of the family of absorbing Markov chains, which can be analysed using the property called time to absorption. The preliminary results show that interesting patterns in student VLE behaviour can be uncovered, especially when combined with the information about submission of the first assessment. Our analysis has been performed using Open University Learning Analytics dataset (OULAD) and research notes are available online (https://bit.ly/2JrY5zv) .

Learning Analytics Dashboard Analysing First-Year Engineering Students

Nowadays, the higher education institutions experience the problem of the student drop-out. In response to this problem, universities started employing analytical dashboards and educational data mining methods such as machine learning, to detect students at risk of failing their studies. In this paper, we present interactive web-based Learning Analytics dashboard - Analyst, which has been successfully deployed at Faculty of Mechanical Engineering (FME), Czech Technical University in Prague. The dashboard provides academic teaching staff with the opportunity to analyse student-related data from various sources in multiple ways to identify those, who might have difficulties to complete their degree. For this purpose, multiple analytical dashboard views have been implemented. It includes summary statistic, study progression graph, and credit completion probabilities graph. In addition, users have the option to export all analysis related graphs for the future use. Based on the outcomes provided by the Analyst, the university successfully ran the interventions on the selected at-risk students and significantly increased the retention rate in the first study year.

Extraction of beats from noisy ECG using ICA

Holter recordings of electrocardiographic signal (ECG) are usually distorted by noise added to measured useful signal due to e. g. worse contact skin-electrode, body movements, etc. Our goal is to create an automatic algorithm for noise removal without destroying morphology of QRS complexes. Our method is based on Independent Component Analysis (ICA). Our so-called ICA “filter” is divided into functional blocks, which can be modified independently on each other. Its performance is tested using two measures, namely Root-Mean-Square error (RMSE) and Pearson’s correlation coefficient.