High-dimensional multi-view dataset pre-processing and clustering

The objective of this project is two-fold: the first one is to perform pre-processing on high-dimensional multi-view datasets and to investigate potential applications of new data mining techniques with the new datasets. The second part of the project is to design and implement an existing...

وصف كامل

محفوظ في:
التفاصيل البيبلوغرافية
المؤلف الرئيسي: Wei, Yaguang.
مؤلفون آخرون: Chen Lihui
التنسيق: Final Year Project
اللغة:English
منشور في: 2012
الموضوعات:
الوصول للمادة أونلاين:http://hdl.handle.net/10356/49881
الوسوم: إضافة وسم
لا توجد وسوم, كن أول من يضع وسما على هذه التسجيلة!
المؤسسة: Nanyang Technological University
اللغة: English
الوصف
الملخص:The objective of this project is two-fold: the first one is to perform pre-processing on high-dimensional multi-view datasets and to investigate potential applications of new data mining techniques with the new datasets. The second part of the project is to design and implement an existing clustering algorithm and conduct extensive experimental study to test the performances of the clustering algorithm on various text benchmark datasets. This report highlights the implementation of the pre-processing approach for the data processing. It also includes the design and implementation ideas of the clustering algorithm Semi-Supervised Spherical K-Means and the simulation results. The performances of SS-SKM are documented and reported.