Title page for etd-0811117-165636


[Back to Results | New Search]

URN etd-0811117-165636
Author Yong-Chun Ding
Author's Email Address No Public.
Statistics This thesis had been viewed 5375 times. Download 0 times.
Department Computer Science and Engineering
Year 2017
Semester 1
Degree Master
Type of Document
Language zh-TW.Big5 Chinese
Title Potential Forecast Algorithm: A Novel Search-Experience-based Clustering Algorithm
Date of Defense 2017-08-28
Page Count 95
Keyword
  • potential forecast algorithm
  • metaheuristic
  • data mining
  • Clustering
  • k-means
  • Abstract Clustering is a classical problem that has been a valuable research topic because it exists in many fields, such as engineering, computer science, medical science, and economics, and it has been widely used as the initial stage in solving these problems. Many algorithms for clustering are likely to fall into local optima easily or are extremely sensitive to the initial solution of the clustering problem, thus making the quality of the end result quite unstable. Therefore, we proposed a search-experience-based algorithm, called potential forecast algorithm (PFA). The underlying idea of the proposed algorithm is to use not only the past searched information to forecast the potential positions which may end up with better solutions, it also uses k-means as a local search mechanism to improve the quality of the end result. To evaluate the performance of PFA, we compare it with other state-of-the-art algorithms. We also test and analyze the influence of all the parameters. The simulation results indicate that PFA can provide not only a better solution but also a more stable quality of the end result.
    Advisory Committee
  • Chung-Nan Lee - chair
  • Chun-Wei Tsai - co-chair
  • Shiann-Rong Kuang - co-chair
  • Ming-Chao Chiang - advisor
  • Files
  • etd-0811117-165636.pdf
  • Indicate in-campus at 99 year and off-campus access at 99 year.
    Date of Submission 2017-09-11

    [Back to Results | New Search]


    Browse | Search All Available ETDs

    If you have more questions or technical problems, please contact eThesys