Wyniki wyszukiwania - BazTech

1

Rule based functional description of genes – Estimation of the multicriteria rule interestingness measure by the UTA method

Gruca A., Sikora M.

Biocybernetics and Biomedical Engineering

|

2013

|

Vol. 33, no. 4

222--234

EN

In this paper we present new extension of RuleGO rule generation method. The method was designed to discover logical rules including combination of GO terms in their premises in order to provide functional description of analyzed gene signatures. As the number of obtained rules is typically huge, filtration algorithm is required to select only the most interesting ones. Rule interestingness measures currently used within the RuleGO method do not always allow for the selection of the rules according to user's subjective preferences. In this paper we propose an application of the UTA method for estimation of the multicriteria rule interestingness measure reflecting expert's subjective rule evaluation. In the presented method, each of the rules is characterized by a vector of values reflecting its quality due to the different parial interestingness measures. From the designated set of rules a set of representative rules is selected and presented to an expert who orders the rules based on his preferences. Using the information about the order and values of the partial interestingness measures, the additive multicriteria interestingness measure is estimated. The measure is estimated in such a way that the rule ranking obtained by this function is consistent with the ranking given by an expert. The presented approach is applied to three microarray data sets and obtained rule orders are compared with rule orders generated with the standard RuleGO rule evaluation method. Presented method allows obtaining the rule ranking that is better correlated with expert ranking than the ranking obtained in the standard way.

2

Modele i reguły decyzyjne w symptomowej diagnostyce technicznej

Jastriebow A., Gad S., Słoń G.

Diagnostyka

|

2006

|

nr 3(39)

199-208

PL

W pracy przedstawiono modele i reguły decyzyjne stosowane w symptomowej diagnostyce technicznej. Modele podzielono na dwie grupy: matematyczne i informacyjne. Opisano kryteria budowy takich modeli. Na podstawie opisanych modeli przedstawiono reguły decyzyjne. Na przykładzie diagnozowania wyposażenia elektrycznego samochodów przedstawiono wyniki symulacyjnej analizy wybranych reguł. Wyniki analizy potwierdzają wysoką efektywność metod opartych na sztucznych sieciach neuronowych.

EN

In the paper models and decision rules, applied in the symptom technical diagnostic, are presented. Models have been split into two groups: mathematical and informative. Criterions of designing of such models are described. On the basis of described models, decision rules have been presented. Results of the simulating analysis of chosen rules have been presented on the example of diagnosing of car's electrical equipment. Results of the analysis confirm high efficiency of methods artificial neural networks.

3

Neural methods of knowledge extraction

Duch W., Adamczak R., Grąbczewski K., Jankowski N.

Control and Cybernetics

|

2000

|

Vol. 29, no 4

997-1017

EN

Contrary to the common opinion, neural networks may be used for knowledge extraction. Recently, a new methodology of logical rule extraction, optimization and application of rule-based systems has been described. C-MLP2LN algorithm, based on constrained multilayer perceptron network, is described here in details and the dynamics of a transition from neural to logical system illustrated. The algorithm handles real-valued features, determining appropriate linguistic variables or membership functions as a part of the rule extraction process. Initial rules are optimized by exploring the accuracy/simplicity tradeoff at the rule extraction stage and the one between reliability of rules and rejection rate at the optimization stage. Gaussian uncertainties of measurements are assumed during application of crisp logical rules, leading to "soft trapezoidal" membership functions and allowing to optimize the linguistic variables using gradient procedures. Comments are made on application of neural networks to knowledge discovery in the benchmark and real life problems.