• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.21 no.2
• /
• pp.534-540
• /
• 1996

This paper presents an Adaptive Time-Delay Recurrent Neural Network (ATRN) for learning and recognition of temporal correlations of temporal patterns. The ATRN employs adaptive time-delays and recurrent connections, which are inspired from neurobiology. In the ATRN, the adaptive time-delays make the ATRN choose the optimal values of time-delays for the temporal location of the important information in the input parrerns, and the recurrent connections enable the network to encode and integrate temporal information of sequences which have arbitrary interval time and arbitrary length of temporal context. The ATRN described in this paper, ATNN proposed by Lin, and TDNN introduced by Waibel were simulated and applied to the chaotic time series preditcion of Mackey-Glass delay-differential equation. The simulation results show that the normalized mean square error (NMSE) of ATRN is 0.0026, while the NMSE values of ATNN and TDNN are 0.014, 0.0117, respectively, and in temporal learning, employing recurrent links in the network is more effective than putting multiple time-delays into the neurons. The best performance is attained bythe ATRN. This ATRN will be sell applicable for temporally continuous domains, such as speech recognition, moving object recognition, motor control, and time-series prediction.

• Journal of Broadcast Engineering
• /
• v.25 no.2
• /
• pp.143-150
• /
• 2020

Huge amounts of contents are being uploaded every day on various streaming platforms. Among those videos, game and sports videos account for a great portion. The broadcasting companies sometimes create and provide highlight videos. However, these tasks are time-consuming and costly. In this paper, we propose models that automatically predict highlights in games and sports matches. While most previous approaches use visual information exclusively, our models use both audio and visual information, and present a way to understand short term and long term flows of videos. We also describe models that combine GAN to find better highlight features. The proposed models are evaluated on e-sports and baseball videos.

• Journal of Korea Water Resources Association
• /
• v.41 no.3
• /
• pp.325-340
• /
• 2008

This study employs Bayesian multiple regression analysis using the ordinary least squares method for regional low flow frequency analysis. The parameter estimates using the Bayesian multiple regression analysis were compared to conventional analysis using the t-distribution. In these comparisons, the mean values from the t-distribution and the Bayesian analysis at each return period are not significantly different. However, the difference between upper and lower limits is remarkably reduced using the Bayesian multiple regression. Therefore, from the point of view of uncertainty analysis, Bayesian multiple regression analysis is more attractive than the conventional method based on a t-distribution because the low flow sample size at the site of interest is typically insufficient to perform low flow frequency analysis. Also, we performed low flow prediction, including confidence interval, at two ungauged catchments in the Nakdong River basin using the developed Bayesian multiple regression model. The Bayesian prediction proves effective to infer the low flow characteristic at the ungauged catchment.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.30B no.12
• /
• pp.59-66
• /
• 1993

In this paper we propose an adaptive update rate tracking algorithm for a phased array radar, based on the interacting multiple model(IMM) algorithm. The purpose of the IMM algorithm hers is twofold: 1) to estimate and predict the target states, and 2) to estimate the level of the process noise. Using the estimate of the process noise level adapted to target dynamics, the update interval is determined to maintain a desired prediction accuracy so that the radar system load is minimized. The adaptive update rate tracking algorithm is implemented for a phased array radar and evaluated with Monte Carlo simulations on various trajectories. The evaluation results of the proposed algorithm and a standard Kalman filter without the adaptive update rate control are presented to compare.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.6
• /
• pp.1957-1980
• /
• 2021

The significant aim of this research has always been to evaluate the mechanism for efficient and inherently aware usage of vitality in-home devices, thus improving the information of smart metering systems with regard to the usage of selected homes and the time of use. Advances in information processing are commonly used to quantify gigantic building activity data steps to boost the activity efficiency of the building energy systems. Here, some smart data mining models are offered to measure, and predict the time series for energy in order to expose different ephemeral principles for using energy. Such considerations illustrate the use of machines in relation to time, such as day hour, time of day, week, month and year relationships within a family unit, which are key components in gathering and separating the effect of consumers behaviors in the use of energy and their pattern of energy prediction. It is necessary to determine the multiple relations through the usage of different appliances from simultaneous information flows. In comparison, specific relations among interval-based instances where multiple appliances use continue for certain duration are difficult to determine. In order to resolve these difficulties, an unsupervised energy time-series data clustering and a frequent pattern mining study as well as a deep learning technique for estimating energy use were presented. A broad test using true data sets that are rich in smart meter data were conducted. The exact results of the appliance designs that were recognized by the proposed model were filled out by Deep Convolutional Neural Networks (CNN) and Recurrent Neural Networks (LSTM and GRU) at each stage, with consolidated accuracy of 94.79%, 97.99%, 99.61%, for 25%, 50%, and 75%, respectively.

• Journal of Korean Society of Transportation
• /
• v.18 no.1
• /
• pp.47-59
• /
• 2000

The purpose of this study is to develop travel-time estimation model using neural networks and prediction model using neural networks and kalman-filtering technique. The data used in this study are travel speed collected from inductive loop vehicle detection systems(VDS) and travel time collected from the toll collection system (TCS) between Seoul and Osan toll Plaza on the Seoul-Pusan Expressway. Two models, one for travel-time estimation and the other for travel-time Prediction were developed. Application cases of each model were divided into two cases, so-called, a single-region and a multiple-region. because of the different characteristics of travel behavior shown on each region. For the evaluation of the travel time estimation and Prediction models, two Parameters. i.e. mode and mean were compared using five-minute interval data sets. The test results show that mode was superior to mean in representing the relationship between speed and travel time. It is, however shown that mean value gives better results in case of insufficient data. It should be noted that the estimation and the Prediction of travel times based on the VDS data have been improved by using neural networks, because the waiting time at exit toll gates can be included for the estimation of travel time based on the VDS data by considering differences between VDS and TCS travel time Patterns in the models. In conclusion, the results show that the developed models decrease estimation and prediction errors. As a result of comparing the developed model with the existing model using the observed data, the equality coefficients of the developed model was average 88% and the existing model was average 68%. Thus, the developed model was improved minimum 17% and maximum 23% rather then existing model .

• Journal of Intelligence and Information Systems
• /
• v.27 no.1
• /
• pp.83-102
• /
• 2021

The government recently announced various policies for developing big-data and artificial intelligence fields to provide a great opportunity to the public with respect to disclosure of high-quality data within public institutions. KSURE(Korea Trade Insurance Corporation) is a major public institution for financial policy in Korea, and thus the company is strongly committed to backing export companies with various systems. Nevertheless, there are still fewer cases of realized business model based on big-data analyses. In this situation, this paper aims to develop a new business model which can be applied to an ex-ante prediction for the likelihood of the insurance accident of credit guarantee. We utilize internal data from KSURE which supports export companies in Korea and apply machine learning models. Then, we conduct performance comparison among the predictive models including Logistic Regression, Random Forest, XGBoost, LightGBM, and DNN(Deep Neural Network). For decades, many researchers have tried to find better models which can help to predict bankruptcy since the ex-ante prediction is crucial for corporate managers, investors, creditors, and other stakeholders. The development of the prediction for financial distress or bankruptcy was originated from Smith(1930), Fitzpatrick(1932), or Merwin(1942). One of the most famous models is the Altman's Z-score model(Altman, 1968) which was based on the multiple discriminant analysis. This model is widely used in both research and practice by this time. The author suggests the score model that utilizes five key financial ratios to predict the probability of bankruptcy in the next two years. Ohlson(1980) introduces logit model to complement some limitations of previous models. Furthermore, Elmer and Borowski(1988) develop and examine a rule-based, automated system which conducts the financial analysis of savings and loans. Since the 1980s, researchers in Korea have started to examine analyses on the prediction of financial distress or bankruptcy. Kim(1987) analyzes financial ratios and develops the prediction model. Also, Han et al.(1995, 1996, 1997, 2003, 2005, 2006) construct the prediction model using various techniques including artificial neural network. Yang(1996) introduces multiple discriminant analysis and logit model. Besides, Kim and Kim(2001) utilize artificial neural network techniques for ex-ante prediction of insolvent enterprises. After that, many scholars have been trying to predict financial distress or bankruptcy more precisely based on diverse models such as Random Forest or SVM. One major distinction of our research from the previous research is that we focus on examining the predicted probability of default for each sample case, not only on investigating the classification accuracy of each model for the entire sample. Most predictive models in this paper show that the level of the accuracy of classification is about 70% based on the entire sample. To be specific, LightGBM model shows the highest accuracy of 71.1% and Logit model indicates the lowest accuracy of 69%. However, we confirm that there are open to multiple interpretations. In the context of the business, we have to put more emphasis on efforts to minimize type 2 error which causes more harmful operating losses for the guaranty company. Thus, we also compare the classification accuracy by splitting predicted probability of the default into ten equal intervals. When we examine the classification accuracy for each interval, Logit model has the highest accuracy of 100% for 0~10% of the predicted probability of the default, however, Logit model has a relatively lower accuracy of 61.5% for 90~100% of the predicted probability of the default. On the other hand, Random Forest, XGBoost, LightGBM, and DNN indicate more desirable results since they indicate a higher level of accuracy for both 0~10% and 90~100% of the predicted probability of the default but have a lower level of accuracy around 50% of the predicted probability of the default. When it comes to the distribution of samples for each predicted probability of the default, both LightGBM and XGBoost models have a relatively large number of samples for both 0~10% and 90~100% of the predicted probability of the default. Although Random Forest model has an advantage with regard to the perspective of classification accuracy with small number of cases, LightGBM or XGBoost could become a more desirable model since they classify large number of cases into the two extreme intervals of the predicted probability of the default, even allowing for their relatively low classification accuracy. Considering the importance of type 2 error and total prediction accuracy, XGBoost and DNN show superior performance. Next, Random Forest and LightGBM show good results, but logistic regression shows the worst performance. However, each predictive model has a comparative advantage in terms of various evaluation standards. For instance, Random Forest model shows almost 100% accuracy for samples which are expected to have a high level of the probability of default. Collectively, we can construct more comprehensive ensemble models which contain multiple classification machine learning models and conduct majority voting for maximizing its overall performance.

• KIPS Transactions on Software and Data Engineering
• /
• v.7 no.9
• /
• pp.351-360
• /
• 2018

The LWR (Locally Weighted Regression) model, which is traditionally a lazy learning model, is designed to obtain the solution of the prediction according to the input variable, the query point, and it is a kind of the regression equation in the short interval obtained as a result of the learning that gives a higher weight value closer to the query point. We study on an incremental ensemble learning approach for LWR, a form of lazy learning and memory-based learning. The proposed incremental ensemble learning method of LWR is to sequentially generate and integrate LWR models over time using a genetic algorithm to obtain a solution of a specific query point. The weaknesses of existing LWR models are that multiple LWR models can be generated based on the indicator function and data sample selection, and the quality of the predictions can also vary depending on this model. However, no research has been conducted to solve the problem of selection or combination of multiple LWR models. In this study, after generating the initial LWR model according to the indicator function and the sample data set, we iterate evolution learning process to obtain the proper indicator function and assess the LWR models applied to the other sample data sets to overcome the data set bias. We adopt Eager learning method to generate and store LWR model gradually when data is generated for all sections. In order to obtain a prediction solution at a specific point in time, an LWR model is generated based on newly generated data within a predetermined interval and then combined with existing LWR models in a section using a genetic algorithm. The proposed method shows better results than the method of selecting multiple LWR models using the simple average method. The results of this study are compared with the predicted results using multiple regression analysis by applying the real data such as the amount of traffic per hour in a specific area and hourly sales of a resting place of the highway, etc.

• Korean Journal of Organic Agriculture
• /
• v.13 no.3
• /
• pp.281-289
• /
• 2005

In order to find out an alternative way of analysis of food waste compost, the Near Infrared Reflectance Spectroscopy(NIRS) was used for the compost assessment because the technics has been known as non-detructive, cost-effective and rapid method. One hundred thirty six compost samples were collected from Incheon food waste compost factory at Namdong Indurial Complex. The samples were analyzed for nitrogen, organic matter (OM), ash, P, and K using Kjedahl, ignition method, and acid extraction with spectrophotometer, respectively. The samples were scanned using FOSS NIRSystem of Model 6500 scanning mono-chromator with wavelength from $400\~2,400nm$ at 2nm interval. Modified partial Least Squares(MPLS) was applied to develop the most reliable calibration model between NIR spectra and sample components such as nitrogen, ash, OM, P, and K. The regression was validated using validation set(n=30). Multiple correlation coefficient($R^2$) and standard error of prediction(SEP) for nitrogen, ash, organic matter, OM/N ratio, P and K were 0.87, 0.06, 0.72, 1.07, 0.68, 1.05, 0.89, 0.31, 0.77, 0.06, and 0.64, 0.07, respectively. The results of this experiment indicates that NIRS is reliable analytical method to assess some components of feed waste compost, also suggests that feasibility of NIRS can be Justified in case of various sample collection around the year.

• Ecology and Resilient Infrastructure
• /
• v.5 no.4
• /
• pp.246-256
• /
• 2018

In order to analyze hydraulic characteristics of discharge coefficient, hydraulic jump height, and hydraulic jump length, accompanied sediment transport, in the under-flow type vertical lift gate, the hydraulic model experiment and dimensional analysis were performed. The correlations between Froude number and hydraulic characteristics were schematized according to the presence and absence of sediment transport; the correlation of hydraulic characteristics and non-dimensional parameters was analyzed and multiple regression formulae were developed. In the hydraulic characteristics accompanied the sediment transport, by identifying the aspect different from the case that the sediment transport is absent, we verified that it is necessary to introduce variables that can express the characteristics of sediment transport. The multiple regression equations were suggested and each determination coefficient appeared high as 0.749 for discharge coefficient, 0.896 for hydraulic jump height, and 0.955 for hydraulic jump length. In order to evaluate the applicability of the developed hydraulic characteristic equations, 95% prediction interval analysis was conducted on the measured and the calculated by regression equations, and it was determined that NSE (Nash-Sutcliffe Efficiency), RMSE (root mean square), and MAPE (mean absolute percentage error) are appropriate, for the accuracy analysis related to the prediction on hydraulic characteristics of discharge coefficient, hydraulic jump height and length.