Use of Dispersion Ratio with Ferrer Diagram for Classification
نام عام مواد
[Thesis]
نام نخستين پديدآور
Shah, Yash Alpeshkumar
نام ساير پديدآوران
Khan, Maleq
وضعیت نشر و پخش و غیره
نام ناشر، پخش کننده و غيره
Texas A&M University - Kingsville
تاریخ نشرو بخش و غیره
2019
يادداشت کلی
متن يادداشت
32 p.
یادداشتهای مربوط به پایان نامه ها
جزئيات پايان نامه و نوع درجه آن
M.S.
کسي که مدرک را اعطا کرده
Texas A&M University - Kingsville
امتياز متن
2019
یادداشتهای مربوط به خلاصه یا چکیده
متن يادداشت
Classification is one of the most useful techniques used in data mining. Classification is used to make predictions by assigning a class label to a given data instance. It is used in various applications like biological classification, document classification, drug discovery, pattern recognition, etc. Data mining is all about extracting hidden information and finding unknown patterns from collected data using various techniques like classification, clustering, and regression. There is a vast amount of data collected worldwide but hardly used for data mining purposes. A lot of research has been done in order to improve the accuracy and efficiency of these techniques. Guggari et al. [1], have suggested using Ferrer diagram feature selection technique to improve accuracy. Roy et al. [2] have suggested using dispersion ratio as splitting criteria instead of Gini index. This thesis combines these two techniques in order to improve the accuracy by using Ferrer diagram technique and replacing Gini index used in Ferrer diagram with dispersion ratio. Rigorous experiments were performed on different datasets regarding heart disease, diabetes, wine quality, and vehicle identification. 80% of the experiments performed show improvement in accuracy ranging between 0.3% - 4.16%.
اصطلاحهای موضوعی کنترل نشده
اصطلاح موضوعی
Computer science
نام شخص به منزله سر شناسه - (مسئولیت معنوی درجه اول )