Backgroung: Type 1 diabetes is a disease that adversely affects the daily life of a large percentage of people worldwide. Daily glucose levels regulation and useful advices provided to patients regarding their diet are essential for diabetes treatment. For this reason, the interest of the academic community has focused on developing innovative systems, such as decision support systems, based on glucose prediction algorithms. The present work presents the predictive capabilities of ensemble methods compared to individual algorithms while combining each method with compartment models for fast acting insulin absorption simulation. Methods: An approach of combining widely used glycemia prediction algorithms is proposed and three different ensemble methods (Linear, Bagging and Boosting metaregressor) are applied and evaluated on their ability to provide accurate predictions for 30, 45 and 60 minutes ahead prediction horizon. Moreover, glycemia levels, long and short acting insulin dosages and consumed carbohydrates from six type one people with diabetes are used as input data and the results are evaluated in terms of root-mean square error and Clarke error grid analysis. Results: According to results, ensemble methods can provide more accurate glucose concentration in comparison to individual algorithms. Bagging metaregressor, specifically, performed better than individual algorithms in all prediction horizons for small datasets. Bagging ensemble method improved the percentage in zone A according to Clarkes error grid analysis by 4% and in some cases by 9%. Moreover, compartment models are proved to improve results in combination with any method at any prediction horizon. This strengthen the potential practical usefulness of the ensemble methods and the importance of building accurate compartment models.
- MeSH
- algoritmy MeSH
- diabetes mellitus 1. typu * farmakoterapie MeSH
- inzulin MeSH
- krevní glukóza MeSH
- lidé MeSH
- selfmonitoring glykemie MeSH
- Check Tag
- lidé MeSH
- Publikační typ
- časopisecké články MeSH
The paper compares two approaches to multi-step ahead glycaemia forecasting. While the direct approach uses a different model for each number of steps ahead, the iterative approach applies one one-step ahead model iteratively. Although it is well known that the iterative approach suffers from the error accumulation problem, there are no clear outcomes supporting a proper choice between those two methods. This paper provides such comparison for different ARX models and shows that the iterative approach outperformed the direct method for one-hour ahead (12-steps ahead) forecasting. Moreover, the classical linear ARX model outperformed more complex non-linear versions for training data covering one-month period.
Cílem příspěvku je ukázat výhody zapojení experta do tvorby znalostního obsahu systému pro podporu rozhodování nejen v klasickém pojetí formalizace znalostí experta do báze znalostí, ale i poněkud méně tradičním způsobem. Tento přístup se začal v nedávné době nazývat „human/expert-in-the-loop“. Základní myšlenkou je využít znalosti experta, které nemusejí být obsaženy v datech, jejichž analýzu chceme provádět, ale které mohou významně zlepšit kvalitu rozhodovacího procesu. Často je možné v rámci tohoto kroku integrovat znalosti více expertů. Pokud se takové přístupy využívají přímo v metodách strojového učení, označují se také jako aktivní učení. Na několika případových studiích z různých oblastí medicíny ukážeme, v jakých fázích procesu vývoje může expert vhodně do procesu zasáhnout. První případová studie je věnována porovnání výsledků získaných pomocí znalostního systému, jehož báze znalostí byla vytvářena manuálně formalizací slovně popsaných znalostí, a pomocí metod strojového učení, konkrétně rozhodovacího stromu, naučeného na větším souboru dat. Následně byla báze znalostí porovnána s rozhodovacím stromem a doplněna o vybraná pravidla z tohoto stromu. Takto upravená báze znalostí poskytovala lepší rozhodnutí než původní. Další dvě případové studie jsou zaměřené na úlohu klasifikace v dlouhodobých záznamech biologických signálů, konkrétně elektroencefalografických a polysomnografických. U této úlohy je klíčové nalézt vhodný poměr mezi zobecněnými metodami, použitelnými na záznamy všech pacientů, a metodami nastavenými na konkrétního pacienta (jedna z možností personalizace). Motivací pro takové řešení je velká interpersonální variabilita, a u řady diagnóz i intrapersonální variabilita. Možnost interaktivního vstupu experta do procesu analýzy záznamů může přispět ke zvýšení kvality a konzistence hodnocení.
This paper aims to show the benefits of expert involvement in the creation of knowledge-based content system to support decision making not only in the classic conception of formalization of expert knowledge into the knowledge base, but also somewhat less traditional way. This approach has recently started to be called "human / expert-in-the-loop". The basic idea is to use the expert knowledge that may not be included in the data, whose analysis we perform, but which can significantly improve the quality of decision making. In this step it is often possible to integrate knowledge of more experts. If such approaches are used in machine learning methods, they are known as active learning. We present several case studies from different areas of medicine, in which we show how experts can interact with the system. The first case study is devoted to compare the results obtained using the knowledge system whose knowledge base was created manually from verbally described expert knowledge, and by using machine learning techniques (specifically the decision tree) learned on a larger data set. Subsequently, the knowledge base was compared with the decision tree and supplemented by selected rules from the tree. This adjusted knowledge base provided better results than the original one. The other two case studies are focused on classification in long-term records of biological signals, namely electroencephalographic and polysomnographic. The key issue is to find the appropriate balance between the generalized methods applicable to the records of all patients, and methods adjusted for each patient (one of the personalization options). The motivation for this approach is the large interpersonal variability, and intrapersonal variability at certain diagnoses. An interactive expert input into the process of analyzing the records can contribute to improvement of the quality and consistency of assessment.
The paper presents a novel encoding scheme for neuronal code generation for odour recognition using an electronic nose (EN). This scheme is based on channel encoding using multiple Gaussian receptive fields superimposed over the temporal EN responses. The encoded data is further applied to a spiking neural network (SNN) for pattern classification. Two forms of SNN, a back-propagation based SpikeProp and a dynamic evolving SNN are used to learn the encoded responses. The effects of information encoding on the performance of SNNs have been investigated. Statistical tests have been performed to determine the contribution of the SNN and the encoding scheme to overall odour discrimination. The approach has been implemented in odour classification of orthodox black tea (Kangra-Himachal Pradesh Region) thereby demonstrating a biomimetic approach for EN data analysis.
- Publikační typ
- abstrakt z konference MeSH
This paper focuses on wrapper-based feature selection for a 1-nearest neighbor classifier. We consider in particular the case of a small sample size with a few hundred instances, which is common in biomedical applications. We propose a technique for calculating the complete bootstrap for a 1-nearest-neighbor classifier (i.e., averaging over all desired test/train partitions of the data). The complete bootstrap and the complete cross-validation error estimate with lower variance are applied as novel selection criteria and are compared with the standard bootstrap and cross-validation in combination with three optimization techniques - sequential forward selection (SFS), binary particle swarm optimization (BPSO) and simplified social impact theory based optimization (SSITO). The experimental comparison based on ten datasets draws the following conclusions: for all three search methods examined here, the complete criteria are a significantly better choice than standard 2-fold cross-validation, 10-fold cross-validation and bootstrap with 50 trials irrespective of the selected output number of iterations. All the complete criterion-based 1NN wrappers with SFS search performed better than the widely-used FILTER and SIMBA methods. We also demonstrate the benefits and properties of our approaches on an important and novel real-world application of automatic detection of the subthalamic nucleus.
- MeSH
- teoretické modely MeSH
- velikost vzorku * MeSH
- Publikační typ
- časopisecké články MeSH
- práce podpořená grantem MeSH
- validační studie MeSH
Complex fractionated atrial electrograms (CFAEs) may represent the electrophysiological substrate for atrial fibrillation (AF). Progress in signal processing algorithms to identify sites of CFAEs is crucial for the development of AF ablation strategies. A novel algorithm for automated description of fractionation of atrial electrograms (A-EGMs) based on the wavelet transform has been proposed. The algorithm was developed and validated using a representative set of 1.5 s A-EGM (n = 113) ranked by three experts into four categories: 1-organized atrial activity; 2-mild; 3-intermediate; 4-high degree of fractionation. A tight relationship between a fractionation index and expert classification of A-EGMs (Spearman correlation rho = 0.87) was documented with a sensitivity of 82% and specificity of 90% for the identification of highly fractionated A-EGMs. This operator-independent description of A-EGM complexity may be easily incorporated into mapping systems to facilitate CFAE identification and to guide AF substrate ablation.
- MeSH
- algoritmy MeSH
- automatizace MeSH
- elektrokardiografie statistika a číselné údaje MeSH
- fibrilace síní patofyziologie MeSH
- interpretace statistických dat MeSH
- katetrizační ablace MeSH
- lidé středního věku MeSH
- lidé MeSH
- pilotní projekty MeSH
- počítačové zpracování signálu MeSH
- senioři MeSH
- Check Tag
- lidé středního věku MeSH
- lidé MeSH
- mužské pohlaví MeSH
- senioři MeSH
- ženské pohlaví MeSH
- Publikační typ
- práce podpořená grantem MeSH