• Korean Journal of Agricultural Science
• /
• v.49 no.2
• /
• pp.193-202
• /
• 2022

The recurrent neural network (RNN) algorithm has been widely used in water-related research areas, such as water level predictions and water quality predictions, due to its excellent time series learning capabilities. However, studies on water quality predictions using RNN algorithms are limited because of the scarcity of water quality data. Therefore, most previous studies related to water quality predictions were based on monthly predictions. In this study, the quality of the water in a reservoir in Nonsan, Chungcheongnam-do Republic of Korea was predicted using the RNN-LSTM algorithm. The study was conducted after constructing data that could then be, linearly interpolated as daily data. In this study, we attempt to predict the water quality on the 7th, 15th, 30th, 45th and 60th days instead of making daily predictions of water quality factors. For daily predictions, linear interpolated daily water quality data and daily weather data (rainfall, average temperature, and average wind speed) were used. The results of predicting water quality concentrations (chemical oxygen demand [COD], dissolved oxygen [DO], suspended solid [SS], total nitrogen [T-N], total phosphorus [TP]) through the LSTM algorithm indicated that the predictive value was high on the 7th and 15th days. In the 30th day predictions, the COD and DO items showed R² that exceeded 0.6 at all points, whereas the SS, T-N, and T-P items showed differences depending on the factor being assessed. In the 45th day predictions, it was found that the accuracy of all water quality predictions except for the DO item was sharply lowered.

• Journal of Biosystems Engineering
• /
• v.35 no.2
• /
• pp.124-131
• /
• 2010

Chlorophyll fluorescence has been researched for assessing fruit post-harvest quality and condition. The objective of this preliminary research was to investigate the potential of fluorescence spectroscopy for measuring apple fruit quality. Ultraviolet (UV) and blue light was used as an excitation source for inducing fluorescence in apples. Fluorescence spectra were measured from 'Golden Delicious' (GD) and 'Red Delicious' (RD) apples using a visible/near-infrared spectrometer after one, three, and five minutes of continuous UV/blue light illumination. Standard destructive tests were performed to measure fruit firmness, skin and flesh color, soluble solids and acid content from the apples. Calibration models for each of the three illumination time periods were developed to predict fruit quality indexes. The results showed that fluorescence emission decreased steadily during the first three minutes of UV/blue light illumination and was stable within five minutes. The differences were minimal in the model prediction results based on fluorescence data at one, three or five minutes of illumination. Overall, better predictions were obtained for apple skin chroma and hue and flesh hue with values for the correlation coefficient of validation between 0.80 and 0.90 for both GD and RD. Relatively poor predictions were obtained for fruit firmness, soluble solids content, titrational acid, and flesh chroma. This research has demonstrated that fluorescence spectroscopy is potentially useful for assessing selected quality attributes of apple fruit and further research is needed to improve fluorescence measurements so that better predictions of fruit quality can be achieved.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.29-29
• /
• 2023

To mitigate the damaging impacts of floods, accurate prediction of runoff, streamflow and flood inundation is needed. Conventional approach of simulating hydrology and hydraulics using loosely coupled models cannot capture the complex dynamics of surface and sub-surface processes. Additionally, the scarcity of data in ungauged basins and quality of data in gauged basins add uncertainty to model predictions, which need to be quantified. In this presentation, first the role of integrated modeling on creating accurate flood simulations and inundation maps will be presented with specific focus on urban environments. Next, the use of machine learning in producing streamflow predictions will be presented with specific focus on incorporating covariate shift and the application of theory guided machine learning. Finally, a framework to quantify the uncertainty in flood models using Hierarchical Bayesian Modeling Averaging will be presented. Overall, this presentation will highlight that creating accurate information on flood magnitude and extent requires innovation and advancement in different aspects related to hydrologic predictions.

• Journal of Soil and Groundwater Environment
• /
• v.21 no.6
• /
• pp.67-79
• /
• 2016

Gaussian process regression (GPR) is proposed as a tool of long-term groundwater quality predictions. The major advantage of GPR is that both prediction and the prediction related uncertainty are provided simultaneously. To demonstrate the applicability of the proposed tool, GPR and a conventional non-parametric trend analysis tool are comparatively applied to synthetic examples. From the application, it has been found that GPR shows better performance compared to the conventional method, especially when the groundwater quality data shows typical non-linear trend. The GPR model is further employed to the long-term groundwater quality predictions based on the data from two domestically operated groundwater monitoring stations. From the applications, it has been shown that the model can make reasonable predictions for the majority of the linear trend cases with a few exceptions of severely non-Gaussian data. Furthermore, for the data shows non-linear trend, GPR with mean of second order equation is successfully applied.

• Journal of Environmental Impact Assessment
• /
• v.9 no.1
• /
• pp.13-24
• /
• 2000

This paper presents a study of water quality model applications in environmental impact statements which were submitted during recent years in Korea. Most of the applications have reported that the development projects would have significant impacts on the water quality, especially, of streams and rivers. The water quality models, however, were hardly used as an impact prediction tool. Even in the cases where models were used, calibration and verification studies were not performed and thus the predicted results would not be reliable. These poor model applications in environmental impact assessment can be attributable to the fact that there were no available model application guidelines as well as no requirements by the review agency. In addition, the expected waste loads were improperly estimated in most cases, especially in non-point sources, and the predicted parameters were not good enough to understand water quality problems expected from the proposed plans. The effects of mitigation measures were not analyzed in most cases. Again, these can be attributed to no formal guidelines available for impact predictions until now. A brief guideline is described in this paper, including model selection, calibration and verification, impact prediction, and analysis of effects of mitigation measures. The results of this study indicate that the model application should be required to overcome the current improper predictions of environmental impacts and the guidelines should be developed in detail and provided.

• Journal of Wetlands Research
• /
• v.10 no.1
• /
• pp.11-20
• /
• 2008

In this study the estimation method for stream hydraulic characteristics which is served as the input data set for running QualKo water quality model is investigated. The conventional approach for estimating such hydraulic parameters is to use the data set from the last cross section in each reach. However, it is shown that in order to represent correctly flow velocity profiles or the travel time in streams, hydraulic parameters of QualKo model should be estimated with all cross section data set within the corresponding reach. In addition, the unsuitable estimation of hydraulic parameters at some reaches has influence on the water quality predictions at the corresponding reaches, and the errors of water quality predictions are propagated toward the downstream without any error attenuation.

• Asian Journal of Atmospheric Environment
• /
• v.8 no.2
• /
• pp.108-114
• /
• 2014

In the companion paper (Lee et al., 2012), it was showed that CMAQ simulation using a lateral boundary conditions (LBCs) derived from RAQMS-CMAQ linkage, compared to the CMAQ results with the default CMAQ LBCs, improved ozone simulations in the conterminous US domain. In the present paper, the study is extended to investigate the influence of LBCs on PM2.5 simulation. MM5-SMOKE-CMAQ modeling system was used for meteorological field generation, emissions preparation and air quality simulations, respectively. Realtime Air Quality Modeling System (RAQMS) model assimilated with satellite observations were used to generate the CMAQ-ready LBCs. CMAQ PM2.5 simulations with RAQMS LBCs and predefined LBCs were compared with U.S. EPA Air Quality System (AQS) measurements. Mean PM2.5 lateral boundary conditions taken from RAQMS outputs showed strong variations both in the horizontal grid and vertical layers in the northern and western boundaries and affected the results of CMAQ PM2.5 predictions. CMAQ with RAQMS LBCs could improve CMAQ PM2.5 predictions resulting in the improvement of index of agreement from 0.38 to 0.63.

• Journal of Korean Society for Atmospheric Environment
• /
• v.32 no.1
• /
• pp.32-45
• /
• 2016

The anthropogenic aerosols originated from the pollutant emissions in the eastern part of China and dust emitted in northwestern China in Yellow sand regions are subsequently transported via eastward wind to the Korean peninsula and then these aerosols induce high $PM_{10}$ concentrations in Korean peninsula. In order to estimate air quality considering anthropogenic and dust emissions, Comprehensive Air-quality Model with extension (CAMx) was applied to simulate $PM_{10}$ concentration. The predicted $PM_{10}$ concentrations without/with dust emissions were compared with observations at ambient air quality monitoring sites in China and Korea for 2008. The predicted $PM_{10}$ concentrations with dust emissions could depict the variation of measured $PM_{10}$ especially during Yellow sand events in Korea. The comparisons also showed that predicted $PM_{10}$ concentrations without dust emissions were under-predicted while predictions of $PM_{10}$ concentrations with dust emission were in good agreement with observations. This implied that dust emissions from desert and barren soil in southern Mongolia and northern China minimized the discrepancies in the $PM_{10}$ predictions in East Asia. The effect of dust emission on annual $PM_{10}$ concentrations in Korea Peninsula for year 2008 was $5{\sim}10{\mu}g/m^3$, which were about 20% of observed annual $PM_{10}$ concentrations.

• Journal of Korean Society for Quality Management
• /
• v.21 no.2
• /
• pp.131-139
• /
• 1993

The newly proposed model of the future reliability results in earlier fault-fixes having a greater effect than the fault which make the greatest contribution to the overall failure rate tend to show themselves earlier, and so are fixed earlier. The suggested model allows a variety of reliability measures to be calculated. Predictions of total execution time(debugging time) is to achieve a target reliability. This model could also apply to computer-hardware reliability growth resulting from the elimination of design error and fault.

• International Journal of Computer Science & Network Security
• /
• v.21 no.2
• /
• pp.148-157
• /
• 2021

The effectiveness of recommendation systems depends on the performance of the algorithms with which these systems are designed. The quality of the algorithms themselves depends on the quality of the strategies with which they were designed. These strategies differ from author to author. Thus, designing a good recommendation system means implementing the good strategies. It's in this context that several research works have been proposed on various strategies applied to algorithms to meet the needs of recommendations. Researchers are trying indefinitely to address this objective of seeking the qualities of recommendation algorithms. In this paper, we propose a new algorithm for recommending learning items. Learner performance predictions and collaborative recommendation methods are used as strategies for this algorithm. The proposed performance prediction model is based on convolutional neural networks (CNN). The results of the performance predictions are used by the proposed recommendation algorithm. The results of the predictions obtained show the efficiency of Deep Learning compared to the k-nearest neighbor (k-NN) algorithm. The proposed recommendation algorithm improves the recommendations of the learners' learning items. This algorithm also has the particularity of dissuading learning items in the learner's profile that are deemed inadequate for his or her training.