"Unsupervised learning is one of the most important steps of machine learning applications. Besides its ability to obtain the insight of the data distribution, unsupervised learning is used as a preprocessing step for other machine learning algorithm. This dissertation investigates the application of unsupervised learning into various types of data for many machine learning tasks such as clustering, regression and classification. The dissertation is organized into three papers. In the first paper, unsupervised learning is applied to mixed categorical and numerical feature data type to transform the data objects from the mixed type feature domain into a new sparser numerical domain. By making use of the data fusion capacity of adaptive resonance theory clustering, the approach is able to reduce the distinction between the numerical and categorical features. The second paper presents a novel method to improve the performance of wind forecast by clustering the time series of the surrounding wind mills into the similar group by using hidden Markov model clustering and using the clustering information to enhance the forecast. A fast forecast method is also introduced by using extreme learning machine which can be trained by analytic form to choose the optimal value of past samples for prediction and appropriate size of the neural network. In the third paper, unsupervised learning is used to automatically learn the feature from the dataset itself without human design of sophisticated feature extractors. The paper points out that by using unsupervised feature learning with multi-quadric radial basis function extreme learning machine the performance of the classifier is better than several other supervised learning methods. The paper further improves the speed of training the neural network by presenting an algorithm that runs parallel on GPU"--Abstract, page iv.
Wunsch, Donald C.
Moss, Randy Hays, 1953-
Stanley, R. Joe
Beetner, Daryl G.
Samaranayake, V. A.
Electrical and Computer Engineering
Ph. D. in Computer Engineering
National Science Foundation (U.S.)
Missouri University of Science and Technology. Intelligent Systems Center
Mary K. Finley Missouri Endowment
Intelligent Systems Center
Research Center/Lab(s) 2 ETDs
Applied Computational Intelligence Lab (ACIL)
Missouri University of Science and Technology
Journal article titles appearing in thesis/dissertation
- Clustering data of mixed categorical and numerical type with unsupervised feature learning
- Hidden Markov model with information criteria clustering and extreme learning machine regression for wind forecasting
- Unsupervised feature learning classification with radial basis function extreme learning machine using graphic processors
x, 71 pages
© 2016 Dao Minh Lam, All rights reserved.
Dissertation - Open Access
Neural networks (Computer science)
Electronic OCLC #
Lam, Dao Minh, "Clustering: Methodology, hybrid systems, visualization, validation and implementation" (2016). Doctoral Dissertations. 2479.