• MALSORI
• /
• no.44
• /
• pp.105-113
• /
• 2002

In Korean, each digit is monosyllable and some pairs are known to have high confusability, causing performance degradation of connected digit recognition systems. To improve the performance, in this paper, we employ various discriminant analyses (DA) including Linear DA (LDA), Weighted Pairwise Scatter LDA WPS-LDA), Heteroscedastic Discriminant Analysis (HDA), and Maximum Likelihood Linear Transformation (MLLT). We also examine several combinations of various DA for additional performance improvement. Experimental results show that applying any DA mentioned above improves the string accuracy, but the amount of improvement of each DA method varies according to the model complexity or number of mixtures per state. Especially, more than 20% of string error reduction is achieved by applying MLLT after WPS-LDA, compared with the baseline system, when class level of DA is defined as a tied state and 1 mixture per state is used.

• Communications for Statistical Applications and Methods
• /
• v.9 no.2
• /
• pp.543-553
• /
• 2002

Principal component analysis is applied to reduce p-dimensions into q-dimensions ( $q {\leq} p$). Any partition of a collection of data points with p and q variables generated by the application of six hierarchical clustering methods is re-classified by discriminant analysis. From the application of discriminant analysis through each hierarchical clustering method, correct classification ratios are obtained. The results illustrate which method is more reasonable in exploratory data analysis.

• Phonetics and Speech Sciences
• /
• v.9 no.4
• /
• pp.27-34
• /
• 2017

The purpose of this work is to examine the English and Korean voiceless plosives from the Buckeye[15] and Seoul[16] corpus in terms of their static spectral characteristics. The plosives were automatically extracted by a Praat script. In order to estimate the percent correctness in the classification of the plosives, discriminant analyses were performed whose trainings were based on four spectral moments, i.e. the center of gravity, variance, skewness and kurtosis as suggested in [6]. Another set of discriminant analyses were performed based on the spectral tilts. In the last set of analyeses, the spectral moments and tilts were both used in the training. Results showed that the correct classification rate did not exceed around 65% in the best case, which suggested that phonetic cues other than the release burst would be necessary including the dynamic spectral aspects and vowel-onset cues.

• The Korean Journal of Applied Statistics
• /
• v.21 no.6
• /
• pp.969-980
• /
• 2008

This paper presents the comparison of four methods, linear, quadratic, canonical and non-parametric discriminant analyses to discriminate the consumers' taste grade with sensory variables, such as tenderness, juiciness, flavor, and overall acceptability based on Consumer Sensory Survey. The classification ability of each method is measured and compared by the resubstitution error rate.

• Proceedings of the Korean Society of Soil and Groundwater Environment Conference
• /
• 2005.04a
• /
• pp.237-241
• /
• 2005

The physical and chemical characteristics of soils in a small watershed were investigated and the effect of geology and land use on soil quality were examined by using multivariate statistical methods, principal components analysis and discriminant analysis. It was considered that the accumulation of salts in the farmland soils indicated by electrical conductivity, contents of cations and anions and pH was caused by fertilizer input during cultivation. The contents of inorganic components are increased as following order: upland > orchard > paddy field > forest. The results of two discriminant analyses using water extractable inorganic components and their ratios by land use were also clearly classified by discriminant function 1 and 2. In discriminant analysis by components, discriminant function 1 indicated the effect of fertilizer application and increased as following order: upland > orchard > paddy field > forest soil.

• Journal of the Korean Data and Information Science Society
• /
• v.19 no.2
• /
• pp.511-518
• /
• 2008

Medical diagnosis systems have been developed to make the knowledge and expertise of human experts more widely available, therefore achieving high-quality diagnosis. In this study, in order to support the diagnosis by the medical diagnosis system, we have preformed medical diagnosis inference three times: first by a neural network with the backpropagation algorithm, secondly by a discriminant analysis with all of the variables, and thirdly by a discriminant analysis with the selected variables. A discussion on comparison of these three methods has been provided.

• Korean journal of applied entomology
• /
• v.32 no.1
• /
• pp.30-41
• /
• 1993

In the present investigation, some multivariate discriminant analyses were done for each population of the native honeybee workers (Apis cerana), which were selected for 15 different localities in spring and 16 in summer form Korea. when the comparison of both seasons for different populations and regardless of seasons were conducted, all the classification results revealed that the differences were significantly prominent. And the length of fore tibia(FTL) was the best contributed character among the 47 morphometric characters used in the analysis.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.21 no.2
• /
• pp.359-365
• /
• 2020

Sasang Constitution data that were classified by the QSCCII (Questionnaire for the Sasang Constitution Classification II) and artery pulse waves of Chon, Guan, and Chuck data measured using an electronic manometer, were obtained from 732 subjects who visited an oriental hospital. The pulse width, peak height, and number of peaks were extracted from the pulse waves as feature variables. Validity and reliability analyses were performed to obtain the feature variables. The feature variables with high validity and reliability were selected as the discriminant variables. The pulse wave data were divided into training and predicting samples by applying a fivefold cross-validation technique. Discriminant analysis was performed for the training sample, and discriminant functions were obtained. The discriminant functions were applied to the predicting sample and the Sasang Constitution was predicted. The accuracy of prediction was estimated by comparing the predicted Sasang Constitution and that obtained by QSCCII. The accuracy of the predicted Sasang Constitution before (after) age and sex calibration was 73.6 % (70.4 %), 68.4 % (84.2 %), and 74.2 % (67.7 %) for Taeumin, Soumin, and Soyangin, respectively, and 72.5 % (73.8 %) in total.

• International Commerce and Information Review
• /
• v.8 no.3
• /
• pp.71-95
• /
• 2006

In recent, T-commerce is widely dispersed as alternative type of commerce. It is forecasted that t-commerce system is used more than e-commerce system. Therefore more and more t-commerce-related industries are also recognizing that t-commerce is a critical business model. It is needed to understand the concept of t-commerce and develop the t-commerce marketing strategy. CEO analyses consumer's behaviors according to the data about buyers and applies the advantage of t-commerce to the communication with customers. This t-commerce system plays an important role in maximizing customer satisfaction and affecting their intention to reuse it. Therefore this paper attempts to identify T-commerce critical success factors and divide between use-intention group and unuse-intention group by taking out a discriminant function by the discriminant analysis. This lays a foundation in developing T-commerce strategy. According to the discriminant function extracted, convenience factor, amusement factor, system quality factor, product perception factor are significant in the sequence of influential degree. However, usefulness factor and speedy connection factor are not significant. In result, the target hitting rate is 77.9% in the first unuse-intention group and it is 95.2% in the second use-intention group. The total discriminant target hitting rate is computed to higher value, 86.55%. The statistic package, SPSS 12.0, is used to survey and analyse data and test the hypothesis. The validity and reliability of variables are verified by both reliability analysis and factor analysis. The discriminant analysis is used to tell the difference between use-intention group and unuse-intention group.

• Korean journal of applied entomology
• /
• v.32 no.1
• /
• pp.42-50
• /
• 1993

Some multivariate discriminant analyses were done for each population of the native honeybee workers (Apis cerana), which were selected for 11 different localities in spring and 12 in summer from Korea. When the comparison for different localities was conducted, the correct assignment was averaged at 91.67% in spring and 88.44% in summer. And for the comparison between the 2 different seasons, it was averaged at 97.58%. Whereas, that regardless of seasons revealed the lowest correct assignment at 70.16%.