دانلود رایگان مقاله الگوریتم ممتیک برای کشف دانش در پیش بینی باینری بر اساس ماشین بردار پشتیبانی

عنوان فارسی
یک الگوریتم ممتیک جدید برای کشف دانش در پیش بینی های باینری و چندکلاسی بر اساس ماشین بردار پشتیبانی
عنوان انگلیسی
A novel memetic algorithm for discovering knowledge in binary and multi class predictions based on support vector machine
صفحات مقاله فارسی
0
صفحات مقاله انگلیسی
16
سال انتشار
2016
نشریه
الزویر - Elsevier
فرمت مقاله انگلیسی
PDF
کد محصول
E2158
رشته های مرتبط با این مقاله
مهندسی کامپیوتر، مهندسی فناوری اطلاعات
گرایش های مرتبط با این مقاله
الگوریتم ها و محاسبات، داده کاوی
مجله
محاسبات کاربردی نرم - Applied Soft Computing
دانشگاه
دانشکده فناوری اطلاعات، تامیل نادو، هند
کلمات کلیدی
داده کاوی، تقسیم بندی ، انتخاب ویژگی، ارزش شپلی، الگوریتم ژنتیک، الگوریتم ممتیک
۰.۰ (بدون امتیاز)
امتیاز دهید
چکیده

ABSTRACT


 In classification, every feature of the data set is an important contributor towards prediction accuracy and affects the model building cost. To extract the priority features for prediction, a suitable feature selector is schemed. This paper proposes a novel memetic based feature selection model named Shapely Value Embedded Genetic Algorithm (SVEGA). The relevance of each feature towards prediction is measured by assembling genetic algorithms with shapely value measures retrieved from SVEGA. The obtained results are then evaluated using Support Vector Machine (SVM) with different kernel configurations on 11 + 11 benchmark datasets (both binary class and multi class). Eventually, a contrasting analysis is done between SVEGA-SVM and other existing feature selection models. The experimental results with the proposed setup provides robust outcome; hence proving it to be an efficient approach for discovering knowledge via feature selection with improved classification accuracy compared to conventional methods.

نتیجه گیری

5. Discussion and conclusion


The experiments conducted in the study successfully demonstrate the beneficial factors of SVM-SVEGA over a range of other feature selection algorithms and prove it to be a promising technique. The results also show that the choices of the kernel function and feature selection technique have a profound effect on the performance of SVM for both binary and multi data classification. The effect of various SVM kernels impacted by the proposed SVEGA scheme is also investigated. Depending on the nature of the problem – binary class or multi class, appropriate kernels may be chosen. The study conducted can be comprehendingly substituted as a complete expert system modelfor deducing faster clinical diagnosis with better accuracy. The future work can test the proposed techniques on datasets from other domains. Further Ant Colony Optimization and Particle Swarm Analysis may be employed when low cost study is required.


بدون دیدگاه