7713

7185

كارشناسي ارشد

نرم افزار

اصفهان: دانشگاه صنعتي اصفهان، دانشكده برق و كامپيوتر

1391

ده،196ص.: نصور،جدول،نمودار

ص.ع.به فارسي و انگليسي

مريم ذكري

22/2/92

محمدعلي منتظري، محمدرضا احمدزاده

مهندسي برق و كامپيوتر

ID7185

به فارسي و انگليسي: قابل رويت در نسخه ديجيتالي

Feature Selection with the proposed Quantom Model Mohammad khosravi m khosravi@ec iut ac ir Date of Submission Januray 2013 Isfahan University of Technology Isfahan Iran Degree M sc Language Farsi Supervisor Maryam Zekri mzekri@cc iut ac ir Abstract In recent years data has become increasingly larger not only in rows i e number of instances but also in columns i e number of features in many applications such as gene selection from micro array data and text automatic categorization the number of features in the raw data ranges from hundreds to tens of thousands High dimensionality brings great difficulty in pattern recognition machine learning and data mining As data reduction is one of the well known technics in data preprocessing With the development of science and the apparent lack of knowledge of the universe and the ability to interpret the phenomena of the universe by using the implications of acquisition These concepts can be used to solving problems in computer science and make a major contribution in this field One of the common areas is using of the quantum mechanics concepts in order to develop efficient algorithms Due to the importance of data reduction With the concepts study quantum mechanics and atomic models the proposed quantum model was used to solve and modeling the feature selection problem In this model features such as electrons are around the nucleus of an atom and are distributed around it and as the electrons move around the nucleus of an atom The best layer for Features around the nucleus of an atom is obtained during the execution of the algorithm and the core features are ignored Supervised clustering using the proposed algorithm will be able to simultaneously create centers of clusters in order to classify a new instance cluster center is as a representative sample of its class These studies have demonstrated ability predicting ability of new instances classes In order to evaluate the performance the model to select attributes the proposed algorithm with the well known algorithms in this area is compared The evaluation results demonstrate a significant superiority of the proposed algorithm and proposed quantum model for the selection of suitable features Keywords Data reduction Quantum feature selection Quantum atomic model Feature weighting Quantum clustering

مريم ذكري

محمدعلي منتظري، محمدرضا احمدزاده