Introduction Of A Method To Diabetes Diagnosis According To Optimum Rules In Fuzzy Systems Based On Combination Of Data Mining Algorithm (d-t), Evolutionary Algorithms (aco) And Artificial Neural Networks (nn)

Abstract

In time diagnosis of diabetes significantly reduces damages and inconveniences of this disease in society. It may be said that one of the most important problems of diagnosis methods of this disease, particularly in early phases, is not to pay attention to proper features in order to diagnose the disease and as a result weakness in disease diagnosis. This research endeavors to introduce a new method for accurate diagnosis of this disease through usage of a combination of artificial intelligent methods such as fuzzy systems for immediate and accurate decision making, Evolutionary Algorithms (ACO 1) for choosing best rules in fuzzy systems, and artificial neural networks for modeling, structure identification, and parameter identification. The proposed system relying on features of database in the form of combination and interaction succeeded in reaching an accuracy of 95.852% which in comparison to current methods on the one hand and to artificial methods in foresaid references on the other hand, has a proper and very faster performance than other intelligent methods and you can see its accuracy and excellence as an intelligent system.

Classification of Diabetes Mellitus using Modified Particle Swarm Optimization and Least Squares Support Vector Machine

“…Some of the records in this data set contains zeros in some features. Cells with zero entry were kept unchanged (neither deleted nor filled) [13]. The input for the Modified-PSO algorithm is total of 768 records.…”

“…Positive cases are the records with Live label and negative ones are with Die label. FP and FN stand for False Positive and False Negative which are the proportion of negative cases that were incorrectly classified as positive and the proportion of positive cases that were incorrectly classified as negative respectively [13]. http://www.ijcttjournal.org…”

“…The results showed a sensitivity of 99.45% for the classifier and specificity of 100% [18]. A classification Algorithm based on Fuzzy systems, Evolutionary Algorithms (ACO) and ANN techniques was proposed [13].…”

“…The proposed Algorithm is compared with different recently classifier Algorithms which were applied on the same database, as shown in TABLE (III). These techniques are multilayer perceptron (AMMLP) [12], Fuzzy, Decision Tree, ACS and ANN [13], 6 different ANN types and AIS technique [7], Also, GA and ANN [9], MLP and other techniques [11] and Different Evolutionary algorithms with different tools [19]. Fig (2) demonstrates the results of the average classification accuracy of the proposed algorithm against the accuracy of other classification algorithms applied on the same data set.…”

Application Of Ant Colony Algorithm And Principal Components Analysis In The Diagnosis Of Lung Cancer

“…The basic idea of this method was proposed by Dorigo [19,20]. An ant colony system includes simple agents (ants) which are working together to generate a output as a single behavior of the entire system; in other words, the agents make a strong system which is able to find high quality answers for issues that have a large search space.…”