| Author | Kirkpong Kiatpanichagij |
| Call Number | AIT Diss. no.ISE-09-03 |
| Subject(s) | Electromyography
|
| Note | A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Engineering, School of Engineering and Technology |
| Publisher | Asian Institute of Technology |
| Series Statement | Dissertation ; no. ISE-09-03 |
| Abstract | This dissertation describes a preprocessing stage for nonlinear classifier used in wavelet
packet transformation (WPT)-based multichannel surface electromyogram (EMG)
classification. The preprocessing stage named sdPCA, which consists of supervised
discretization coupled with principal component analysis (PCA), was developed for
improving surface EMG classifier generalization ability and training speed on overlap
segmented signals.
The operating principle of supervised discretization is different from dimensionality
reduction, i.e. the size of a feature vector is not reduced. Supervised discretization
partitions the features range using class randomness and converts feature values of the
samples in a partition to integer numbers representing partition order. Usually, it is utilized
in data mining algorithms and discrete classifiers. For example, supervised discretization is
a prerequisite for Naïve Bayes classifier, which converts the continuous features to the
discrete features before the discrete features are fed into a probability model. In addition to
this, its binary version is used in a tree construction phase of a Classification And
Regression Tree (CART). It is also a compulsory preprocessing stage of the fast
correlation-based filter (FCBF), which is a feature section-based dimensionality reduction.
The literature indicates that this dissertation firstly uses supervised discretization as a
preprocessing stage in WPT-based surface EMG classification. The experiments confirm
the impressive performance of supervised discretization compared with other
preprocessing stages. Coupling it with PCA reduces the effects of the curse of
dimensionality and further enhances the total performance. The resultant preprocessing
stage is termed as sdPCA.
The sdPCA outperforms FCBF, PCA, supervised discretization, and their combinations in
terms of the highest generalization ability, fast training speed, the small feature size, and an
ability to reduce the risks of developing oscillation and being trapped in nonlinear
classifier training. The experiments were conducted on a data set consisting of 4-channel
surface EMG signals measured from 6 hand and wrist gestures of 12 subjects. The
experimental results indicate that the classification system using sdPCA has the highest
generalization ability along with the second fastest training speed. The classification
accuracy in 12 subjects of the system using sdPCA is 93.30 ± 2.42% taking 400 epochs for
training by overlap segmented signals within 100 s. This result is very attractive for further
development because high-classification accuracy for large data sets is achieved by means
of the proposed sdPCA without the application of additional algorithms such as local
discriminant bases (LDB), majority voting (MV), or WPT sub-bands clustering. |
| Year | 2009 |
| Corresponding Series Added Entry | Asian Institute of Technology. Dissertation ; no. ISE-09-03 |
| Type | Dissertation |
| School | School of Engineering and Technology (SET) |
| Department | Department of Industrial Systems Engineering (DISE) |
| Academic Program/FoS | Industrial Systems Engineering (ISE) |
| Chairperson(s) | Afzulpurkar, Nitin V.; |
| Examination Committee(s) | Manukid Parnichkun ;Guha, Sumanta; |
| Scholarship Donor(s) | Royal Thai Government; |
| Degree | Thesis (Ph.D.) - Asian Institute of Technology, 2009 |