• ETRI Journal
• /
• v.36 no.5
• /
• pp.714-720
• /
• 2014

In this paper, we propose the use of data-driven probabilistic utterance-level decision logic to improve Weighted Finite State Transducer (WFST)-based endpoint detection. In general, endpoint detection is dealt with using two cascaded decision processes. The first process is frame-level speech/non-speech classification based on statistical hypothesis testing, and the second process is a heuristic-knowledge-based utterance-level speech boundary decision. To handle these two processes within a unified framework, we propose a WFST-based approach. However, a WFST-based approach has the same limitations as conventional approaches in that the utterance-level decision is based on heuristic knowledge and the decision parameters are tuned sequentially. Therefore, to obtain decision knowledge from a speech corpus and optimize the parameters at the same time, we propose the use of data-driven probabilistic utterance-level decision logic. The proposed method reduces the average detection failure rate by about 14% for various noisy-speech corpora collected for an endpoint detection evaluation.

• Journal of Global Scholars of Marketing Science
• /
• v.18 no.2
• /
• pp.47-64
• /
• 2008

It is well known that consumer's quality perception is strongly affected by price. Higher priced products tend to be perceived to have better quality than lower priced products although the objective product quality is the same. However, it is less known the process through which quality perception is affected by price cues. The existing literature suggests three potential hypotheses (i.e., the selectiveprocessing hypothesis, the selective interpretation hypothesis, and the representativeness heuristic hypothesis) that explain the underlying processes of the price-induced quality perception. The current research tests among the three competing hypotheses and also examines the role of consumer knowledge as a moderating factor. An experiment was conducted to test the moderating role of the knowledge in the price-quality relationship and to investigate the underlying process. The results indicate that theinfluence of price on perceived quality differs between novices and experts. Expert consumers' quality perception is not significantly influenced by price and this finding is consistent with the extant literature. On the other hand, novice consumers' quality perception is affected by price and the tests on the underlying process support for the representativeness heuristic hypothesis. Novice consumers assess that a high-priced brand should have good product quality due to the belief that high (low) price brands represent better (worse) brand quality and such a representativeness heuristic occurs without involving selective attention or selective interpretation price-consistent information.

• Research in Mathematical Education
• /
• v.15 no.4
• /
• pp.311-325
• /
• 2011

In this paper I propose that teaching students the most efficient method of problem solving may curtail students' creativity. Instead it is important to arm students with a variety of problem solving heuristics. It is the students' responsibility to decide which heuristic will solve the problem. The chosen heuristic is the one which is meaningful to the students.

• Proceedings of the IEEK Conference
• /
• 2001.06c
• /
• pp.59-62
• /
• 2001

This paper discusses a data mining technique for time series analysis of traffic data, which provides useful knowledge for network configuration management. Commonly, a network designer must employ a combination of heuristic algorithms and analysis in an interactive manner until satisfactory solutions are obtained. The problem of heuristic algorithms is that it is difficult to deal with large networks and simplification or assumptions have to be made to make them solvable. Various data mining techniques are studied to gain valuable knowledge in large and complex telecommunication networks. In this paper, we propose a traffic pattern association technique among network nodes, which produces association rules of traffic fluctuation patterns among network nodes. Discovered rules can be utilized for improving network topologies and dynamic routing performance.

• Journal of Korean Institute of Industrial Engineers
• /
• v.23 no.1
• /
• pp.95-113
• /
• 1997

This paper outlines a framework for a diagnosis of a complex system with uncertain information. Sensor validation ploys a vital role in the ability of the overall system to correctly determine the state of a system monitored by imperfect sensors. Here, emphases are put on the heuristic technology and post-processor for reasoning. Heuristic Sensor Validation (HSV) exploits deeper knowledge about parameter interaction within the plant to cull sensor faults from the data stream. Finally the modified probability distributions and validated data are used as input to the reasoning scheme which is the runtime version of the influence diagram. The output of the influence diagram is a diagnostic mapping from the symptoms or sensor readings to a determination of likely failure modes. Once likely failure modes are identified, a detailed diagnostic knowledge base suggests corrective actions to improve performance. This framework for a diagnostic expert system with sensor validation and reasoning under uncertainty applies in $HEATXPRT^{TM}$ a data-driven on-line expert system for diagnosing heat rate degradation problems in fossil power plants [1].

• Research in Mathematical Education
• /
• v.25 no.2
• /
• pp.135-158
• /
• 2022

This study aimed to building models to understand the relationships between reasoning resources and content knowledge. We applied Support Vector Machine and linear models to the data including fifth graders' scores in the Cornel Critical Thinking Test and the Iowa Assessments, demographic information, and learning science approach (a student-centered approach to learning called the Science Writing Heuristic [SWH] or traditional). The SWH model showing the relationships between critical thinking domains and academic achievement at grade 5 was developed, and its validity was tested across different learning environments. We also evaluated the stability of the model by applying the SWH models to the data of the grade levels. The findings can help mathematics educators understand how critical thinking and achievement relate to each other. Furthermore, the findings suggested that reasoning in mathematics classrooms can promote performance on standardized tests.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 1996.11a
• /
• pp.588-592
• /
• 1996

In this paper, we have constructed an expert system for semiconductor chip encapsulation which combines a knowledge-based system with CAE software. The knowledge-base module includes heuristic and pre-analysis knowledge for evaluation and redesign. Evaluation of the initial design and generation of redesign recommendations can be developed from the rules as applied to a given chip Package. The CAE programs can be used for simulating the filling and packing stage of encapsulation process. The expert system is a new tool which enables package design or process conditions with high yields and high productivity.

• Journal of the Korea Safety Management & Science
• /
• v.2 no.2
• /
• pp.1-10
• /
• 2000

This study deals with the application of knowledge engineering and a methodology for the assessment and measurement of reliability, availability, maintainability, and safety of industrial systems using fault-tree representation. A fuzzy methodology for fault-tree evaluation seems to be an alternative solution to overcome the drawbacks of the conventional approach (insufficient information concerning the relative frequence of hazard events). To improve the quality of results, the membership functions must be approximated based on heuristic considerations. The purpose of this study is to describe the knowledge engineering approach, directed to integrate the various sources of knowledge involved in a FTA.

• Journal of Information Processing Systems
• /
• v.16 no.2
• /
• pp.478-493
• /
• 2020

We designed a new meta-heuristic algorithm named Photon Search Algorithm (PSA) in this paper, which is motivated by photon properties in the field of physics. The physical knowledge involved in this paper includes three main concepts: Principle of Constancy of Light Velocity, Uncertainty Principle and Pauli Exclusion Principle. Based on these physical knowledges, we developed mathematical formulations and models of the proposed algorithm. Moreover, in order to confirm the convergence capability of the algorithm proposed, we compared it with 7 unimodal benchmark functions and 23 multimodal benchmark functions. Experimental results indicate that PSA has better global convergence and higher searching efficiency. Although the performance of the algorithm in solving the optimal solution of certain functions is slightly inferior to that of the existing heuristic algorithm, it is better than the existing algorithm in solving most functions. On balance, PSA has relatively better convergence performance than the existing metaheuristic algorithms.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.17 no.3
• /
• pp.87-94
• /
• 2003

When faults occur in a distribution network, blackout loads must be restored as fast as possible to minimize interruption of electric service. This paper presents an expert system to restore distribution networks. The system is designed to minimize switching operations in a heuristic sense by using heuristic rules. The restoration process is converted to the feasible path finding problem in the state space of expert system by application of general topological knowledge base. Multiple load transfer algorithm is proposed to deal with complex situations.