naser nematbakhsh

Diabetes is a group of metabolic disorders that are the result of untreated high blood glucose greatly. Early diagnosis and continued control of this disease can reduce its effectsTherefore, providing a method for timely diagnosis of this disease is of great importance. Until now, researchers have made many efforts to provide machine learning methods to diagnose diabetes. But most of these models are either based on simple machine learning methods or based on the assumption that the available diabetes data are balanced. Both cases are factors of their complete failure. Therefore, considering the existing challenges as well as the importance of timely diagnosis of diabetes, in this research, a diabetes diagnosis model based on deep recurrent neural networks and SMOTE oversampling algorithm is presented. In this model, several pre-processing steps including quantification of missing values, removal of outlier data and then oversampling have been performed. Three deep recurrent neural networks with three recurrent hidden units including LSTM, GRU and BiLSTM have been used to diagnose diabetes. The results of the model presented on the Pima database received from the UCI repository indicate that the average accuracy in 10 different runs in LSTM, GRU and BiLSTM is 91.21%, 89.61% and 90.99%, respectively. The recurrent network with GRU unit has achieved the highest accuracy of 93.74% on average in 10 different executions. The results of the proposed model show that deep neural networks have a much more successful performance in diabetes diagnosis compared to traditional machine learning methods.

In general, Policy-based approaches play an important role in the management of web services, for instance, in the choice of semantic web service and quality of services (QoS) in particular. The present research work illustrates a procedure for the web service selection among functionality similar web services based on WS-Policy semantic matching. In this study, the procedure of WS-Policy publishing in the UDDI registry was also described. The approach, which is used to represent the policies, is thus represented as semantic trees, and in this representation, measurable quality attributes are considered; and the certain matching operations are used to identify the similarity match via match function or similarity distance function. The illustration of semantic concepts and rules during policy matching, which is not possible by using a mere semantic concept, leads to better web service matches. The proposed approach has been validated through various tests that can evaluate the similarity of large and arbitrary sets of measurable quality attributes. We also compared the proposed procedure with the other ones. The proposed procedure for web service choose, which uses WS-Policy semantic matching, can be more effective to solve different problems like selection, composition, and substitution of services.

Data mining has been used as a public utility in extracting knowledge from databases during recent years. Developments in data mining and availability of data and private information are the biggest challenge in this regard. Privacy preserving data mining is a response to this big challenge. The main purpose of techniques and algorithms in privacy preserving data mining is non-disclosure of sensitive and private data with minimum changes in databases so that it would not have adverse effects on the rest of data. The present paper intends to present a brief review of methods and techniques regarding privacy of data mining in association rules, their classification and finally, classification of hiding algorithms of association rules followed by a comparison between a numbers of these algorithms.

Software testing is one of the most important phases in the software development procedure which ensures the accordance of the software and its description. Testing is mainly a manual task accomplished by the human operators. This results in increasing the cost and time of the software development process. Also, due to the uncertain nature of the human activities, software reliability will be under threat and the probability of having some aspects and parts of the software untested always would be high. Therefore, the more automatic, the more intelligent, and the more reliable testing procedure always would be of interest. In this paper we introduce a new approach to the software testing automation in web based applications, using Artificial Neural Network (ANN). The applied ANN will be trained by diverse pairs of input/output data provided according to the software functionality, then it attempts to model a testing tool for the software. Next we can use this ANN-based testing tool to evaluate and test the software. We apply the proposed testing scheme on a modified version of a web based university course registration software and show its performance on both error-free and faulty cases.