• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.7 no.7
• /
• pp.1657-1670
• /
• 2013

The increase in the number of music files in smart phone and MP3 player makes it difficult to find the music files which people want. So, Query-by-Singing/Humming (QbSH) systems have been developed to retrieve music from a user's humming or singing without having to know detailed information about the title or singer of song. Most previous researches on QbSH have been conducted using musical instrument digital interface (MIDI) files as reference songs. However, the production of MIDI files is a time-consuming process. In addition, more and more music files are newly published with the development of music market. Consequently, the method of using the more common MPEG-1 audio layer 3 (MP3) files for reference songs is considered as an alternative. However, there is little previous research on QbSH with MP3 files because an MP3 file has a different waveform due to background music and multiple (polyphonic) melodies compared to the humming/singing query. To overcome these problems, we propose a new QbSH method using MP3 files on mobile device. This research is novel in four ways. First, this is the first research on QbSH using MP3 files as reference songs. Second, the start and end positions on the MP3 file to be matched are estimated by using multi-layered perceptron (MLP) prior to performing the matching with humming/singing query file. Third, for more accurate results, four MLPs are used, which produce the start and end positions for dynamic time warping (DTW) matching algorithm, and those for chroma-based DTW algorithm, respectively. Fourth, two matching scores by the DTW and chroma-based DTW algorithms are combined by using PRODUCT rule, through which a higher matching accuracy is obtained. Experimental results with AFA MP3 database show that the accuracy (Top 1 accuracy of 98%, with an MRR of 0.989) of the proposed method is much higher than that of other methods. We also showed the effectiveness of the proposed system on consumer mobile device.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.22 no.4
• /
• pp.311-318
• /
• 2016

The VTS concept was located as a principal methods of maritime safety administration in world's major harbors and expected to become the pivotal role for the future of the maritime and harbor society with e-Navigation epoch. If recent limelight concept of big-data has been included in aspect of information gathering and analysis with various studies, it's required advanced studies to improve the information analysis capability and application range of the data that can be mining by the VTS. In this study, contrast to other studies that aimed quantitative analysis as communication number, it can be mining the time information and each of the communication VTS for the target vessel, including qualitative analysis, such as the purpose or the type of communication. This comparison across multiple items of the collected information, and presenting the VTS data mining model (VCDF) that can be analyzed for the purpose of analyzing way, type and number of communication by ship's type, also number of violations through VTS communication. First, In Busan port case, it shows frequently information service and shows frequently communicating with particular types of vessels. Second, Passive VTS carried out notwithstanding many kinds of traffic violations due to communication congestion. This arranged information can be used as data for the analysis, as possible the level of traffic for VTSO situational awareness, which pointed to the 'workloads' in 'IALA Guideline' and could be used as a database for future research of e-Navigation.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.18 no.8
• /
• pp.41-47
• /
• 2017

With the introduction of new technologies, ground weapons have led to the development of artificial intelligence and the attention of major developed countries. In this study, CFD was performed through the BLU-103 model to obtain aerodynamic data for aircraft that are subjected to rotational motion. To simulate the steady-state of a rotating body, the body was fixed and the principle of rotating the body by rotating the surrounding air was used. In order to examine the aerodynamic feasibility of the rotating aircraft, the analysis was carried out at intervals of $30^{\circ}$ angle from $0^{\circ}$ to $90^{\circ}$ for the simple shape and the side slip angle. It was confirmed that the drag coefficient for the simple model satisfies the quantitative results of 1.0 ~ 1.2 through CD presented in "Drag Book". The aerodynamic data was constructed by applying the valid input verified through the simple type analysis conditions to the actual shape, and the tendency was analyzed. The analysis confirmed that CX, CZ and CY increase not only in the simple model but also in the rotation of the actual model. Especially, the influence of CZ was judged to have contributed to the flight.

• Journal of KIISE:Software and Applications
• /
• v.31 no.12
• /
• pp.1684-1690
• /
• 2004

This paper presents a method for classifier selection that uses distribution information of the training samples in a small region surrounding a sample. The conventional DCS-LA(Dynamic Classifier Selection - Local Accuracy) selects a classifier dynamically by comparing the local accuracy of each classifier at the test time, which inevitably requires long classification time. On the other hand, in the proposed approach, the best classifier in a local region is stored in the FSA(Feature Space Attribute) table during the training time, and the test is done by just referring to the table. Therefore, this approach enables fast classification because classification is not needed during test. Two feature space attributes are used entropy and density of k training samples around each sample. Each sample in the feature space is mapped into a point in the attribute space made by two attributes. The attribute space is divided into regular rectangular cells in which the local accuracy of each classifier is appended. The cells with associated local accuracy comprise the FSA table. During test, when a test sample is applied, the cell to which the test sample belongs is determined first by calculating the two attributes, and then, the most accurate classifier is chosen from the FSA table. To show the effectiveness of the proposed algorithm, it is compared with the conventional DCS -LA using the Elena database. The experiments show that the accuracy of the proposed algorithm is almost same as DCS-LA, but the classification time is about four times faster than that.

• Steel and Composite Structures
• /
• v.35 no.6
• /
• pp.799-813
• /
• 2020

Fatigue cracks of rib-to-deck (RD) joints have been frequently observed in the orthotropic steel decks (OSD) using conventional U-ribs (CU). Thickened edge U-rib (TEU) is proposed to enhance the fatigue strength of RD joints, and its effectiveness has been proved through fatigue tests. In-depth full-scale tests are further carried out to investigate both the fatigue strength and fractography of RD joints. Based on the test result, the mean fatigue strength of TEU specimens is 21% and 17% higher than that of CU specimens in terms of nominal and hot spot stress, respectively. Meanwhile, the development of fatigue cracks has been measured using the strain gauges installed along the welded joint. It is found that such the crack remains almost in semi-elliptical shape during the initiation and propagation. For the further application of TEUs, the design curve under the specific survival rate is required for the RD joints using TEUs. Since the fatigue strength of welded joints is highly scattered, the design curves derived by using the limited test data only are not reliable enough to be used as the reference. On this ground, an experiment-numerical hybrid approach is employed. Basing on the fatigue test, a probabilistic assessment model has been established to predict the fatigue strength of RD joints. In the model, the randomness in material properties, initial flaws and local geometries has been taken into consideration. The multiple-site initiation and coalescence of fatigue cracks are also considered to improve the accuracy. Validation of the model has been rigorously conducted using the test data. By extending the validated model, large-scale databases of fatigue life could be generated in a short period. Through the regression analysis on the generated database, design curves of the RD joint have been derived under the 95% survival rate. As the result, FAT 85 and FAT 110 curves with the power index m of 2.89 are recommended in the fatigue evaluation on the RD joint using TEUs in terms of nominal stress and hot spot stress respectively. Meanwhile, FAT 70 and FAT 90 curves with m of 2.92 are suggested in the evaluation on the RD joint using CUs in terms of nominal stress and hot spot stress, respectively.

• Journal of KIISE:Databases
• /
• v.30 no.3
• /
• pp.336-349
• /
• 2003

Histograms have been getting a lot of attention recently. Histograms are commonly utilized in commercial database systems to capture attribute value distributions for query optimization Recently, in the advent of researches on approximate query answering and stream data, the interests in histograms are widely being spread. The simplest approach assumes that the attributes in relational tables are independent by AVI(Attribute Value Independence) assumption. However, this assumption is not generally valid for real-life datasets. To alleviate the problem of approximation on multi-dimensional data with multiple one-dimensional histograms, several techniques such as wavelet, random sampling and multi-dimensional histograms are proposed. Among them, GENHIST is a multi-dimensional histogram that is designed to approximate the data distribution with real attributes. It uses overlapping buckets that allow more efficient approximation on the data distribution. In this paper, we propose a scheme, OPT that can determine the optimal frequencies of overlapped buckets that minimize the SSE(Sum Squared Error). A histogram with overlapping buckets is first generated by GENHIST and OPT can improve the histogram by calculating the optimal frequency for each bucket. Our experimental result confirms that our technique can improve the accuracy of histograms generated by GENHIST significantly.

• Journal of Service Research and Studies
• /
• v.3 no.2
• /
• pp.53-60
• /
• 2013

Datamining from the large data in the form of various techniques for obtaining information have been developed. In recent years one of the most sought areas of pattern recognition and machine learning method is created with most of existing learning algorithms based on categorical attributes to a rule or decision model. However, the real-world data, it may consist of numeric attributes in many cases. In addition it contains attributes with numerical values to the normal categorical attribute. In this case, therefore, it is required processes in order to use the data to learn an appropriate value for the type attribute. In this paper, the domain of the numeric attributes are divided into several segments using learning algorithm techniques of discritization. It is described Clustering with other data mining techniques. Large amount of first cluster with characteristics is similar records from the database into smaller groups that split multiple given finite patterns in the pattern space. It is close to each other of a set of patterns that together make up a bunch. Among the set without specifying a particular category in a given data by extracting a pattern. It will be described similar grouping of data clustering technique to classify the data.

• Asian Pacific Journal of Cancer Prevention
• /
• v.16 no.7
• /
• pp.2833-2838
• /
• 2015

Background: Repeating a prior chemotherapy (rechallenge therapy) is an option for selected patients with metastatic colorectal cancer, but there is very little evidence in the literature for this approach. Thus, we reviewed our registry to evaluate prognostic factors and survival of patients who received irinotecan and oxaliplatin-based regimens as rechallenge third and fourth-line therapy. Materials and Methods: Patients who received irinotecan-based or oxaliplatin-base regimen as first-line had been rechallenged with third-line or fourth-line therapy. These patients were selected from the database of Turkish mCRC registry archives between October 2006 and June 2013 and evaluated retrospectively for factors effecting progression free survival (PFS) and overall survival (OS) by the Kaplan-Meire and Cox-regression methods. Results: Thirty-nine patients were enrolled. The median duration of follow-up was 36 months (14-68 months). Thirty-one patients (76%) died during follow-up. In terms of rechallenge treatments, 29 patients had received third-line and 10 patients had received fourth-line. Response rate (RR) was found to be 12.9%, with stable disease in 19 (48.7%) patients. The median PFS was 6 months (95%CI=4.64-7.35 months) and the median OS was 11 months (95%CI=8.31-13.68 months). The factors effecting survival (PFS and OS) were only being PFS after first-line chemotherapy ${\geq}12months$ (p=0.007, 95% CI=1.75-35.22 and p=0.004, 95%CI=1.44-7.11), both in univariate and multivariate analyses. Conclusions: This study indicates that rechallenge treatment could be a good option as a third or later line therapy in patients who had ${\geq}12months$ PFS onreceiving first line therapy.

• Korea Journal of Hospital Management
• /
• v.5 no.1
• /
• pp.176-199
• /
• 2000

This study is to identify factors affecting public relations(PR) activities through newspapers among hospitals. 53 acute hospitals with more than fill beds were surveyed by telephone interview, and hospital-related articles on 4 major daily newspapers were searched in KINDS(Korea Intergrated Newspapers Database System). Hospitals' PR scores were evaluated by weighting the directoin and the number of hospitals mentioned in an articles. The scores were analysed as the outcome of PR activities by some general characteristics. The mean of PR score was 4.64, most hospitals(22, 41.5%) fell on the group scored more than 0 and less than 5. Hospitals with score under zero were 7(13.2%) and the hospitals scored over 20 point were 3(5.8%). The number of beds and area, two general characteristics of hospitals, had decisive effect on PR activities. Analysis showed hospitals over 1000 beds scored 17.06 point and this group accounted for the highest. Hospitals from 600 to 799 beds scored 2.86, with 800 to 999beds scored 1.93 and with 500 to 599 beds scored 0.59. Along with area, hospitals in Seoul scored 11.50 point, in Kyong-Ki and Incheon scored 1.50 and in the other areas scored -0.23 point. Department for PR and the number of workers in PR activities made hospitals active in PR. Hospitals with PR department scored 7.05 point, without PR department scored 0.98. Hospitals with workers who is whole responsibility over 5 scored 13.67, from 3 to 4 scored 6.54, from 1 to 2 scored 1.02, and the hospitals without workers who is whole responsibility scored 0.53. Among the characteristics related with PR activities, the assignment of a report in individual hospital had significant effect on the outcome of PR. Hospitals with assigned reporters scored 8.31 point, without them scored 1.89. Finally, the factors affecting PR score were identified by multiple regression analysis. The number of beds and the area(Seoul) were significantly associated with the score in positive direction.

• Journal of muscle and joint health
• /
• v.24 no.3
• /
• pp.227-237
• /
• 2017

Purpose: The purpose of this study is to analyze the suitability and applicability of the recommendations for the exercise intervention program in fall prevention guidelines for the elderly. Methods: We searched systematically RISS, KISS, National Assembly Library, KoreaMed, and KM base for Korean guidelines and used OVID-MEDLINE, EMBASE, Cochrane library, Trip database, GIN, NGC, WHO, and CDC for international guidelines including fall prevention exercises for the elderly until August 2016. Results: A total of 11 guidelines were selected as the result of the literature search. After evaluating the quality of guidelines, 9 guidelines were used to analyze the recommendations. A total of 19 recommendations were derived from 9 guidelines. Of 19 recommendations, 12 recommendations were considered to be appropriate and applicable. As a result, we recommend for the elderly to do a single type of balance exercise, strength exercise, aerobic exercise or multiple forms of exercise for 20~30 minutes per a day, 2~3 days or more than 3 days per a week, and consistently for more than 10 weeks. Conclusion: To prevent falls in the elderly, it is important for the elderly to perform a regular balance, strength, or aerobic exercise for 20~30 minutes per a day, 2~3 days per a week.