The goal of research is the fall detection in elderly residents based on infra red depth sensor measurements. Our attention is focused on statistical properties as generalization. The effectiveness of discriminative statistical classifiers (multilayer perceptron) is improved by addition of feature selection block by Gram-Schmidt orthogonalization, which determines the ranking of the features, and NPCA block, which transforms the raw data into a nonlinear manifold and reduces the dimensionality of the data. Performance of our system measured in terms of sensitivity is 92% and precision is 93%, which means it can be used for real life applications.
The paper presents a new approach to computer-aided analysis of ECG Holter recordings. In contrast to existing tools it is a learning system: the pertinent features of the signal shape are automatically discovered upon the examples carefully selected and commented by cardiologists. Mathematical basis of our system is the theory of support vector machines that are applied for two tasks: signal approximation and pattern classification. Numerical procedures implement the algorithm of sequential minimal optimisation. The computer program is developed in Borland C++ Builder environment. The excellent performances of our approach, high rate of successful pattern recognition and computational efficiency, make use of our tools possible in clinical practice. The system is tested at the Chair and Department of Internal Medicine and Cardiology, Central Teaching Hospital in Warsaw, Poland.
The paper describes results of research on the fall detection in elderly residents based on infra red depth sensor measurements. We present the methodology of data acquisition, preprocessing and the feature extraction. Multilayer perceptron is used for classification. In order to improve the classifier generalization feature selection block by Gram-Schmidt orthogonalization is added. It determines the ranking of the features and enables to reduce the dimensionality of the data. Performance of our system measured in terms of sensitivity is 92% and precision is 93%, which means it can be used for real life applications.
We present the improved k-nearest neighbor (kNN) classifier and its application for biomedical data. Our method limits the number of considered neighbors from the training set by selecting only those samples that are neighbors in the computed Voronoi diagram of the training set plus classified sample. The method is based on convex hull calculation of inversed set of points. A very important feature of presented method is the stability of results (in terms of recall and precision values) in broad range of the neighborhood size as opposed to the regular kNN classifier. The classification performance was confirmed on three biomedical benchmark data sets.
This article summarises the results of implementation of a Graph Neural Network classi er. The Graph Neural Network model is a connectionist model, capable of processing various types of structured data, including non- positional and cyclic graphs. In order to operate correctly, the GNN model must implement a transition function being a contraction map, which is assured by imposing a penalty on model weights. This article presents research results concerning the impact of the penalty parameter on the model training process and the practical decisions that were made during the GNN implementation process.
Using the adequate number of Multilayer Perceptron input, hidden and output layers, the Cellular Neural Network dynamic behavior, when the system converges to a fixed-point, can be reproduced by a Multilayer Perceptron with restrictions. A Multilayer Perceptron can then be defined in order to act as a two neuron Cellular Neural Network and vice-versa. From this, we combine their properties in order to overcome the CNN learning problem.