• Journal of Wetlands Research
• /
• v.25 no.4
• /
• pp.306-314
• /
• 2023

Stream sediments are an important component of water quality management because they are receptors of various pollutants such as heavy metals and organic matters emitted from upland sources and can be secondary pollution sources, adversely affecting water environment. To effectively manage the stream sediments, identification of primary sources of sediment contamination and source-associated control strategies will be required. We evaluated the performance of machine learning models in identifying primary sources of sediment contamination based on the physico-chemical properties of stream sediments. A total of 356 stream sediment data sets of 18 quality parameters including 10 heavy metal species(Cd, Cu, Pb, Ni, As, Zn, Cr, Hg, Li, and Al), 3 soil parameters(clay, silt, and sand fractions), and 5 water quality parameters(water content, loss on ignition, total organic carbon, total nitrogen, and total phosphorous) were collected near abandoned metal mines and industrial complexes across the four major river basins in Korea. Two machine learning algorithms, linear discriminant analysis (LDA) and support vector machine (SVM) classifiers were used to classify the sediments into four cases of different combinations of the sampling period and locations (i.e., mine in dry season, mine in wet season, industrial complex in dry season, and　industrial complex in wet season). Both models showed good performance in the classification, with SVM outperformed LDA; the accuracy values of LDA and SVM were 79.5% and 88.1%, respectively. An SVM ensemble model was used for multi-label classification of the multiple contamination sources inlcuding landuses in the upland areas within 1 km radius from the sampling sites. The results showed that the multi-label classifier was comparable performance with sinlgle-label SVM in classifying mines and industrial complexes, but was less accurate in classifying dominant land uses (50~60%). The poor performance of the multi-label SVM is likely due to the overfitting caused by small data sets compared to the complexity of the model. A larger data set might increase the performance of the machine learning models in identifying contamination sources.

• Korean Journal of Biological Psychiatry
• /
• v.5 no.1
• /
• pp.95-101
• /
• 1998

The changes of electroencephalogram(EEG) in patients with dementia of Alzheimer's type are most commonly studied by analyzing power or magnitude in traditionally defined frequency bands. However because of the absence of an identified metric which quantifies the complex amount of information, there are many limitations in using such a linear method. According to the chaos theory, irregular signals of EEG can be also resulted from low dimensional deterministic chaos. Chaotic nonlinear dynamics in the EEG can be studied by calculating the largest Lyapunov exponent($L_1$). The authors have analyzed EEG epochs from three patients with dementia of Alzheimer's type and three matched control subjects. The largest $L_1$ is calculated from EEG epochs consisting of 16,384 data points per channel in 15 channels. The results showed that patients with dementia of Alzheimer's type had significantly lower $L_1$ than non-demented controls on 8 channels. Topographic analysis showed that the $L_1$ were significantly lower in patients with Alzheimer's disease on all the frontal, temporal, central, and occipital head regions. These results show that brains of patients with dementia of Alzheimer's type have a decreased chaotic quality of electrophysiological behavior. We conclude that the nonlinear analysis such as calculating the $L_1$ can be a promising tool for detecting relative changes in the complexity of brain dynamics.

• Tuberculosis and Respiratory Diseases
• /
• v.56 no.4
• /
• pp.393-404
• /
• 2004

Background : Nucleic acid hybridization has become an essential technique in the development of our understanding of gene structure and function. The quantitative analysis of hybridization has been used in the measurement of genome complexity and gene copy number. The filter hybridization assay is rapid, sensitive and can be used to measure RNAs complementary to any cloned DNA sequence. Methods : The authors assessed the accuracy, linearity, correlation coefficient and specificity of the hybridization depending on the added dose(0, 1, 5, and $10{\mu}g$) of non-specific rat spleen RNA to hybridization of surfactant protein A mRNA. Filter hybridization assays were used to obtain the equation of standard curve and thereby to quantitate the mRNA quantitation. Results : 1. Standard curve equation of filter hybridization assay between counts per minute (X) and spleen RNA input (Y) was Y=0.13X-19.35. Correlation coefficient was 0.98. 2. Standard curve equation of filter hybridization assay between counts per minute (X) and surfactant protein A mRNA transcript input (Y) was Y=0.00066X-0.046. Correlation coefficient was 0.99. 3. Standard curve equation of filter hybridization assay between counts per minute (X) and surfactant protein A mRNA transcript input (Y) after the addition of $1{\mu}g$ spleen RNA was Y=0.00056X-0.051. Correlation coefficient was 0.99. 4. Standard curve equation of filter hybridization assay between counts per minute (X) and surfactant protein A mRNA transcript input (Y) after the addition of $5{\mu}g$ spleen RNA was Y=0.00065X-0.088. Correlation coefficient was 0.99. 5. Standard curve equation of filter hybridization assay between counts per minute (X) and surfactant protein A mRNA transcript input (Y) after the addition of $10{\mu}g$ spleen RNA was Y=0.00051X-0.10. Correlation coefficient was 0.99. Conclusions : Comparison of cpm/filter in a linear range allowed accurate and reproducible estimation of surfactant protein A mRNA copy number irrespective of the addition dosage of non-specific rat spleen RNA over the range $0-10{\mu}g$.