A5064933175- Statistical Distribution Estimation and Applications
- Probabilistic and Robust Engineering Design
- Educational Technology and Assessment
- Reliability and Maintenance Optimization
- Laser Design and Applications
- Hydrological Forecasting Using AI
- Online Learning and Analytics
- Advanced Statistical Methods and Models
- High voltage insulation and dielectric phenomena
- Hydrology and Drought Analysis
- Energy Load and Power Forecasting
- Neural Networks and Applications
- Spectroscopy and Laser Applications
- Solid State Laser Technologies
- Fault Detection and Control Systems
- Power System Reliability and Maintenance
- Non-Destructive Testing Techniques
- Face and Expression Recognition
- Machine Learning and Data Classification
- Power Transformer Diagnostics and Insulation
- Laser-induced spectroscopy and plasma
- Statistical Methods and Inference
- COVID-19 epidemiological studies
- Infrastructure Maintenance and Monitoring
- Intelligent Tutoring Systems and Adaptive Learning
Kurume University
2020-2023
Hiroshima Institute of Technology
2015-2019
Warren Clinic
2018-2019
Kaneka (Japan)
2017
Kyushu Institute of Technology
2006-2015
Tokyo University of Technology
2015
Tohoku University
2011
Kyushu University
1983-2010
Jichi Medical University
2010
Hiroshima City University
1995-2007
In the present article, an attempt is made to derive optimal data-driven machine learning methods for forecasting average daily and monthly rainfall of Fukuoka city in Japan. This comparative study conducted concentrating on three aspects: modelling inputs, pre-processing techniques. A comparison between linear correlation analysis mutual information find input technique. For rainfall, a novel hybrid multi-model method proposed compared with its constituent models. The models include...
This paper describes a new statistical method of lifetime estimation solid electrical insulation using data from an accelerated aging test. The simultaneously uses the maximum likelihood and Newton-Raphson methods. novel aspect this work is introduction threshold term in addition to well-known techniques based on inverse power model combined with Weibull distribution. Thus, we can estimate value confidence interval. In order reduce time costs voltage-lifetime test5, testing procedure type I...
Nowadays, detecting the disaster phenomena and predicting final stage become very important in risk analysis view-point. The statistical methods provide accurate estimates of parameters when data are completely given. However, incomplete, accuracy becomes poor. Therefore, weak future trends. SIR methods, for infectious disease spread prediction, using differential equations can sometimes stage. These however, require some inspection time, which means delay at least one week or so we want to...
The author presents a method that uses accelerated life-test data to estimate the mean life at service stress and threshold below which failure is unlikely occur. relation between mean-life assumed follow an inverse power law includes stress. times given are Weibull distribution in shape parameter varies with This model extends well-known model. If only but not specific percentile point sought, maximum likelihood useful for estimation. tradeoff parametric approach. For adoption of...
The main contribution of this paper is to provide reasonable confidence intervals for maximum likelihood estimates percentile points in dielectric breakdown voltage probability distributions. Weibull distributions which include threshold values are often used In some samples, there exist cases the all three parameters diverge; these cases, limiting forms, correspond Gumbel However, computing point using observed information matrix might be impossible when distribution models assumed type....
For original article by H. Hirose, see ibid., vol.3, p.43-55, 1996. This paper discusses the maximum likelihood estimation of a three-parameter Weibull distribution through generalized extreme value (GEV) given cumulative function. The purpose this correspondence is not just to point out shortcomings Hirose. We feel references as well results mentioned can help readers and authors journal in making appropriate choices with regard similar modeling problems
This paper provides a simple bias correction method for the maximum likelihood estimates in 2-parameter Weibull distribution. The unbiased of shape and scale parameters as well percentile points can be obtained by using formula.
To diagnose the electric power apparatus, decision tree method can be a highly recommended classification tool because it provides if-then-rule in visible, and thus we may have possibility to connect physical phenomena observed signals. The most important point constructing diagnosing system is make clear relations between faults corresponding Such database built up laboratory using model made it. next thing feature extraction. We used oslash - V n patterns POW for variables, extraction by...
To predict the continuous value of target variable using values explanation variables, we often use multiple linear regression methods, and many applications have been successfully reported. However, in some data cases, methods may not work because strong local dependency to variables. In such k nearest-neighbor method (k-NN) can be an alternative. Although a simple k-NN improves prediction accuracy, newly proposed method, combined (NNRMLR), is found show accuracy improvement. The NNRMLR...
The item response theory (IRT) gives us the valuable information about difficulties of problems as well abilities students, whereas classical test method provides only students with pre-determined scores to each problem. To enhance use IRT, we have developed a concise IRT evaluation Web system via drag-and-drop Excel file in which 0/1 result are stored. In addition, introduced an online adaptive assess students' more accurately fewer problems. such system, bank is pre-stored and problem...
Pandemic simulation is considered to be crucial as a scenario and it performed by many kinds of methods, the classical ordinary differential models (SIR model), agent-based models, internet-based etc are among them. The SIR model one fundamental methods see behavior pandemic with easy computation. However, there no stochastic variation in equations. equations (SDE) can provide such kind variations. Although SDE applied fields economics, less attention has been paid simulations. In this...
Due to the shrinkage of high school students and growth number acceptance universities in Japan, almost all are be enrolled universities, which causes various educational difficulties universities. Taking care who were taught insufficiently schools junior is one them. The follow-up program aimed at helping need basic learning assisting teachers have engage teaching a variety students. systems recognized as part program, consist check testing, collaborative work testing. They connected...
In applying the adaptive testing equipped with item response theory to actual cases, it is crucial know appropriate number of question items suffice for specified accuracy. almost all systems we have developed so far, adopt small in a sequence questions because too many will bore examinees or force give up completing tests although estimation accuracy can be obtained. By experience, set five. However, cause less accurate estimates examinees' abilities. this paper, show how influences...
The IEC, which prescribes lightning impulse test techniques for electric power equipment, is about to introduce the k-factor method evaluating overshoot of waveforms. Its procedure specifies that waveforms are be fitted with a double exponential function derive base curve. Some derived waveforms, however, deviate upwards from central lines recorded in wavefront area. This paper examines more rational extracting curve by considering relationships between equivalent circuits and solutions...
Rainfall forecasting has been one of the most scientifically and technologically challenging task in climate dynamics prediction theory around world last century. This is due to great effect on human activities also for significant computational advances that are utilized this research field. In paper our main objective forecast over a very short-term specified local area weather using data which not always available by center but will be future social network or some other methods. For...
To take care of students who were taught insufficiently in high schools and junior schools, we have recently established the follow-up program aimed at helping need basic learning assisting teachers to engage teaching a variety educational students. The systems are recognized as part program, consist check testing, collaborative work testing. These testing use large number structured problem items installed database (i.e., item bank). In this paper, introduce system configuration show how...