Taoyuan, Taiwan

Time filter

Source Type

Sun H.-M.,Kainan University
Pattern Recognition | Year: 2010

Up-to-date skin detection techniques use adaptive skin color modeling to overcome the varying skin color problem. Most methods for tracking skin regions in videos utilize the correlation between contiguous frames. This paper proposes a new approach for detecting skin in a single image. This approach uses a local skin model to shift a globally trained skin model to adapt the final skin model to the current image. Experimental results show that the proposed method can achieve better accuracy. Two improvements for speeding up the processing are also discussed. © 2009 Elsevier Ltd. All rights reserved.


Chung Y.-S.,Kainan University
Accident Analysis and Prevention | Year: 2013

Factor complexity is a characteristic of traffic crashes. This paper proposes a novel method, namely boosted regression trees (BRT), to investigate the complex and nonlinear relationships in high-variance traffic crash data. The Taiwanese 2004-2005 single-vehicle motorcycle crash data are used to demonstrate the utility of BRT. Traditional logistic regression and classification and regression tree (CART) models are also used to compare their estimation results and external validities. Both the in-sample cross-validation and out-of-sample validation results show that an increase in tree complexity provides improved, although declining, classification performance, indicating a limited factor complexity of single-vehicle motorcycle crashes. The effects of crucial variables including geographical, time, and sociodemographic factors explain some fatal crashes. Relatively unique fatal crashes are better approximated by interactive terms, especially combinations of behavioral factors. BRT models generally provide improved transferability than conventional logistic regression and CART models. This study also discusses the implications of the results for devising safety policies. © 2012 Elsevier Ltd.


While the previous literature shows that a decline in energy intensity represents an improvement in energy use efficiency, it does not provide a target level of energy intensity, nor what room for improvement in terms of energy intensity could entail. This study establishes an indicator of such room for improvement in terms of energy intensity by measuring the difference between the target level of energy intensity and the actual energy intensity and thereby monitors energy use efficiency. The traditional indicator of energy intensity, defined as energy use over GDP, mainly estimates energy use efficiency, but is a partial effect between the energy input and GDP output. However, our proposed indicator of the room for improvement in terms of energy intensity is the total-factor effects based on the multiple-inputs model. By taking the 27 EU members to investigate their energy use efficiency using the indicator of the room for improvement in terms of energy intensity, this study concludes that an improvement in energy intensity does not fully depend on a decline in energy intensity, and we instead need to confirm whether the room for improvement in terms of energy intensity decreases. This finding is particularly relevant for energy policy-makers. © 2013 Elsevier Ltd.


Chang M.-C.,Kainan University
Energy Policy | Year: 2013

This study provides a no-output growth model to conveniently calculate the total-factor energy efficiency (TFEE) index originally proposed by Hu and Wang (2006). The TFEE index serves as a very well-known and popular means of estimating overall energy efficiency. While many previous studies have used the indicator of energy inefficiency, including the indicator of energy intensity (i.e., Energy input/Gross Domestic Product (GDP)) to measure energy efficiency, Hu and Kao (2007) point out that the indicator of energy intensity is not only a partial-factor energy efficiency indicator, but that this partial-factor ratio is also quite inappropriate for analyzing the impact of changing energy use over time. The TFEE index overcomes the disadvantage of the indicator of energy intensity as mentioned above, but five steps are needed to calculate the TFEE score. In this study, we provide a no-output growth model to conveniently calculate the TFEE score. Furthermore, we extend this no-output growth model to an output growth model. This study concludes that the output growth model not only makes it easier to calculate the TFEE index than the model proposed by Hu and Wang (2006) and Hu and Kao (2007), but that it can also obtain better TFEE scores. © 2012 Elsevier Ltd.


Pai C.-W.,Kainan University
Accident Analysis and Prevention | Year: 2011

Recent emphasis on bicycling as an alternative to automobile transportation has underscored the need for research efforts directed at bicycle safety when sharing roadways with motorised vehicles. Much of the research attention is focused on junction accidents where motorists tend to infringe upon bicycles' right of way. Non-junction accidents where a motorist strikes a bicycle while overtaking it, or crashes into the rear of the bicycle, have been less frequently researched. Another common crash type is a door crash that involves a bicycle striking an open door of an automobile. Using British Stats19 accident data, the present study estimates a mixed multinomial model to predict the likelihood of a non-junction crash being of a certain crash type (out of three possible types). The methodological approach adopted allows for the individuals within the observations to have different parameter estimates (as opposed to a single parameter representing all observations). Main findings include that buses/coaches as collision partners were associated with overtaking crashes; and bicycles' traversing manoeuvres were associated with overtaking and rear-end collisions. Given a crash where a bicycle collides with a motorcycle/taxi, it is more likely a rear-end crash and a door crash, respectively. Implications of the research findings, the concluding remarks, and recommendations for future research are finally provided. © 2011 Elsevier Ltd All rights reserved.


Pai C.-W.,Kainan University
Accident Analysis and Prevention | Year: 2011

The most typical automobile-motorcycle collision take places when an automobile manoeuvres into the path of an approaching motorcycle by violating the motorcycle's right of way (ROW). Aim: The present paper provides a comprehensive review of past research that examined motorcycle ROW accidents. Methods: Articles and publications were selected for relevance and research strength through a comprehensive search of major databases such as Transportation Research Information Services (TRIS), Compendex, and Medline. Results: Two major causes of such a crash scenario are the lack of motorcycle conspicuity and motorist's speed/distance judgment error, respectively. A substantial number of studies have manipulated physical characteristics of motorcycles and motorcyclists to enhance conspicuity, along with research addressing motorists' gap-acceptance behaviours and arrival time judgments when confronting motorcycles. Although various conspicuity aids have proven effective, some researchers reported that motorcyclist's/motorcycle's brightness per se may be less important as a determinant of conspicuity than brightness contrast between the motorcyclists and the surroundings. Larger vehicles tended to be judged to arrive sooner than motorcycles. Such a speed/distance judgment error is likely attributable to some psychological effects such that larger automobiles appear more threatening than motorcycles. Older motorists particularly have difficulties in accurately estimating the distance and the speed of an approaching motorcycle. Research examining the effects of conspicuity measures on motorists' speed/distance judgments when confronting motorcycles has been rather inconclusive. Conclusions: Past research offers valuable insight into the underlying motorcycle ROW crash mechanisms. However, with ageing society and a rapid change in traffic composition (e.g.; more larger motorcycles) in recent years, prior research findings should be updated. The present study finally provides recommendations for future research on motorcycle ROW accidents. © 2010 Elsevier Ltd All rights reserved.


Technology-Enabled Active Learning (TEAL) is a pedagogical innovation established in a technology-enhanced multimedia studio, emphasizing constructivist-oriented teaching and learning. In Taiwan, an increasing number of schools are adopting the TEAL notion to deliver courses. This study examines the impact of TEAL on both student performance and teachers' teaching of physics in the context of one of the high schools. A quasi-experimental research approach was used to conduct the study. Data sources include pre-/post-tests, interviews, class observations.; the researcher's journals. The findings reveal that the benefits that the participants gained from exposure to the innovative instruction appear in various aspects in addition to the students' test results. Having higher interest in attending physics classes and being more active in participating in extracurricular science activities on the part of the students.; being more enthusiastic about and confident in helping students strengthen their physics concepts on the part of the teacher, are among the non-test score gains. The students' achievements and positive responses to the teacher's instruction seem to have motivated his dedication and confidence. It is also found that teachers' teaching beliefs and desire to change greatly affected their classroom practices and technology integration. To more effectively implement instructional innovations in a school, suggestions are provided. © 2012 Elsevier Ltd. All rights reserved.


Chuang C.-L.,Kainan University
Artificial Intelligence in Medicine | Year: 2011

Objectives: In Taiwan, as well as in the other countries around the world, liver disease has reigned over the list of leading causes of mortality, and its resistance to early detection renders the disease even more threatening. It is therefore crucial to develop an auxiliary system for diagnosing liver disease so as to enhance the efficiency of medical diagnosis and to expedite the delivery of proper medical treatment. Methods: The study accordingly integrated the case-based reasoning (CBR) model into several common classification methods of data mining techniques, including back-propagation neural network (BPN), classification and regression tree, logistic regression, and discriminatory analysis, in an attempt to develop a more efficient model for early diagnosis of liver disease and to enhance classification accuracy. To minimize possible bias, this study used a ten-fold cross-validation to select a best model for more precise diagnosis results and to reduce problems caused by false diagnosis. Results: Through a comparison of five single models, BPN and CBR emerged to be the top two methods in terms of overall performance. For enhancing diagnosis performance, CBR was integrated with other methods, and the results indicated that the accuracy and sensitivity of each CBR-added hybrid model were higher than those of each single model. Of all the CBR-added hybrid models, the BPN-CBR method took the lead in terms of diagnosis capacity with an accuracy rate of 95%, a sensitivity of 98%, and a specificity of 94%. Conclusions: After comparing the five single and hybrid models, the study found BPN-CBR the best model capable of helping physicians to determine the existence of liver disease, achieve an accurate diagnosis, diminish the possibility of a false diagnosis being given to sick people, and avoid the delay of clinical treatment. © 2011 Elsevier B.V.


Tseng F.-C.,Kainan University
Expert Systems with Applications | Year: 2013

Although many methods have been proposed to enhance the efficiencies of data mining, little research has been devoted to the issue of scalability-that is, the problem of mining frequent itemsets when the size of the database is very large. This study proposes a methodology, hierarchical partitioning, for mining frequent itemsets in large databases, based on a novel data structure called the Frequent Pattern List (FPL). One of the major features of the FPL is its ability to partition the database, and thus transform the database into a set of sub-databases of manageable sizes. As a result, a divide-And-conquer approach can be developed to perform the desired data-mining tasks. Experimental results show that hierarchical partitioning is capable of mining frequent itemsets and frequent closed itemsets in very large databases. © 2012 Elsevier Ltd. All rights reserved.


Chuang C.-L.,Kainan University
Information Sciences | Year: 2013

Predicting business failure is an important and challenging issue that has served as an impetus for many academic studies over the past three decades. This study aims at developing CBR-based hybrid models of predicting business failure. The need to supplement CBR (Case-Based Reasoning) with other classification and diagnosis techniques is triggered by the fact that accuracy and effectiveness tend to get reduced when CBR alone is applied to handle too many attributes. To enhance the accuracy of bankruptcy prediction, the hybrid models developed by this study include: RST-CBR (combining Rough Set Theory with CBR), RST-GRA-CBR (integrating RST, Grey Relational Analysis, and CBR), and CART-CBR (combining Classification and Regression Tree with CBR). In order to verify the ability of the proposed models to achieve optimal accuracy rate, this study further compares the predictive ability of CBR with those of other comparative models. © 2013 Elsevier Inc. All rights reserved.

Loading Kainan University collaborators
Loading Kainan University collaborators