In Distributed industrial control systems it is necessary to guarantee certain reliability level. In this sense, Checkpointing and Rollback techniques offer interesting possibilities to achieve fault tolerance without appreciable cost and complexity increment. Several Checkpointing techniques have been proposed. Most of them suppose the presence of stable storage in the system. But distributed industrial control systems usually do not dispose of this kind of storage. So, another storage strategy has to be employed. If Checkpoints were locally stored (Simple Checkpointing), the system tolerates only transient faults. If Checkpoints were locally, at the same node, and, additionally, at another/s node/s of the system stored (Two-level Checkpointing), the system can recover from some permanent faults too. In this article the results of a study of the reliability of these two different Checkpoint storage strategies were presented in order to evaluate if the reliability increase of the Two-level method justifies its greater complexity. In order to accomplish this study, two distributed industrial control systems were presented. Each of them are based on a different node architecture which will have an important effect upon the results of the study.
This article presents a method of modeling based on a data set by means of the GMDH algorithms. The data set contains information about fresh students of the Faculty of Electrical Engineering, Czech Technical University of Prague, Czech Republic (CTU, FEE). The motivation for our investigation was to discover, whether and how a particular student will be successful in his/her university studies. The data set was mined from official study application forms. They reflect student's personality, type and results (grades) from high school and entrance examination. These results serve as input vectors for prediction of his/her study results after the 1st semester. The results produced by the model were then compared with the real ones. The second motivation for our investigation was to find the significance of a particular input vector component, because it enables us to identify possible weak points of a student. As appropriate tools the GMDH neural networks (both linear and nonlinear) have been used.
This article reports a new approach to rule extraction method by using Group Method of Data Handling (GMDH) Algorithm in Data Mining area. The advantages of this method are (1) it accepts both categorical and continuous data at the same time, and (2) rules can be extracted easily from the generated model.We applied GMDH Algorithm to categorical data set of US congress voting records to extract rules. The correction rate of GMDH rules was 97.3% -- higher than Tsukimoto's method of rule extraction from Back-propagation neural network (81.0%). It was also higher than 97.0% of C4.5.
Three self-organizing data mining technologies that employ complementary descriptive languages - parametric regression models (GMDH neural networks), fuzzy rules (self-organizing fuzzy rule induction), and similarity models (analog complexing based clustering and classification) - are applied to generate diagnosis models of different levels of heart disease. The classification results show an accuracy of over 95% in average. Due to the strong knowledge extraction capabilities of the used technologies a nucleus of 4 most relevant variables is identified. The obtained results both classification accuracy and identified nucleus are also important for diagnosis cost reduction considerations.
