Auscultating Diagnosis for Hemodialysis Shunt Stenosis using a Self-Organizing Map and Hidden Markov Model

Bearing performance degradation assessment is essential to avoid abrupt machinery breakdown. However, background noise, outliers, and other interferences in the monitoring data may restrict the accuracy and stability of bearing performance degradation assessment in practical applications. In this study, a bearing performance degradation assessment method based on the topological representation and hidden Markov model is proposed. To construct a robust and representative feature space, the topological representations, specifically, topological meshes of the original features are obtained by self-organizing map, which can represent the general structure of the original feature space and eliminate outliers and other interferences. Then, the weight vectors of topological meshes are used as degradation features. Finally, the hidden Markov model is adopted as the assessment model to evaluate the bearing performance degradation tendency and detect the initial degradation effectively. To validate the effectiveness and superiority of the proposed method, two experimental datasets are analyzed. Compared with peer methods, the performance indicator curve of the proposed method presents a more smooth and accurate degradation tendency than comparative methods. Moreover, initial degradation can be identified accurately.

Download Full-text

Probability Model Based on Cluster Analysis to Classify Sequences of Observations for Small Training Sets

Statistics Optimization & Information Computing ◽

10.19139/soic-2310-5070-690 ◽

2020 ◽

Vol 8 (1) ◽

pp. 296-303

Author(s):

Sergey S Yulin ◽

Irina N Palamar

Keyword(s):

Markov Chain ◽

Markov Model ◽

Hidden Markov Model ◽

Short Term Memory ◽

Hidden Markov ◽

Probability Model ◽

Training Data ◽

Training Dataset ◽

Experimental Comparison ◽

Self Organizing Map

The problem of recognizing patterns, when there are few training data available, is particularly relevant and arises in cases when collection of training data is expensive or essentially impossible. The work proposes a new probability model MC&CL (Markov Chain and Clusters) based on a combination of markov chain and algorithm of clustering (self-organizing map of Kohonen, k-means method), to solve a problem of classifying sequences of observations, when the amount of training dataset is low. An original experimental comparison is made between the developed model (MC&CL) and a number of the other popular models to classify sequences: HMM (Hidden Markov Model), HCRF (Hidden Conditional Random Fields),LSTM (Long Short-Term Memory), kNN+DTW (k-Nearest Neighbors algorithm + Dynamic Time Warping algorithm). A comparison is made using synthetic random sequences, generated from the hidden markov model, with noise added to training specimens. The best accuracy of classifying the suggested model is shown, as compared to those under review, when the amount of training data is low.

Download Full-text