• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.6 no.12
• /
• pp.3197-3218
• /
• 2012

Compared with traditional itinerary planning, intention-oriented itinerary recommendations can provide more flexible activity planning without requiring the user's predetermined destinations and is especially helpful for those in unfamiliar environments. The rank and classification of points of interest (POI) from location-based social networks (LBSN) are used to indicate different user intentions. The mining of vehicles' physical trajectories can provide exact civil traffic information for path planning. This paper proposes a POI category-based itinerary recommendation framework combining physical trajectories with LBSN. Specifically, a Voronoi graph-based GPS trajectory analysis method is utilized to build traffic information networks, and an ant colony algorithm for multi-object optimization is implemented to locate the most appropriate itineraries. We conduct experiments on datasets from the Foursquare and GeoLife projects. A test of users' satisfaction with the recommended items is also performed. Our results show that the satisfaction level reaches an average of 80%.

• Journal of Computing Science and Engineering
• /
• v.3 no.3
• /
• pp.165-180
• /
• 2009

The HKIB, or Hankookilbo, test collections are two archives of Korean newswire stories manually categorized with semi-hierarchical or hierarchical category taxonomies. The base newswire stories were made available by the Hankook Ilbo (The Korea Daily) for research purposes. At first, Chungnam National University and KISTI collaborated to manually tag 40,075 news stories with categories by semi-hierarchical and balanced three-level classification scheme, where each news story has only one level-3 category (single-labeling). We refer to this original data set as HKIB-40075 test collection. And then Yonsei University and KISTI collaborated to select 20,000 newswire stories from the HKIB-40075 test collection, to rearrange the classification scheme to be fully hierarchical but unbalanced, and to assign one or more categories to each news story (multi-labeling). We refer to this modified data set as HKIB-20000 test collection. We benchmark a k-NN categorization algorithm both on HKIB-20000 and on HKIB-40075, illustrating properties of the collections, providing baseline results for future studies, and suggesting new directions for further research on Korean text categorization problem.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.05a
• /
• pp.408-410
• /
• 2021

In the NLP field, the pre-training model BERT launched by the Google team in 2018 has shown amazing results in various tasks in the NLP field. Subsequently, many variant models have been derived based on the original BERT, such as RoBERTa, ERNIEBERT and so on. In this paper, the WWMBERT (Whole Word Masking BERT) model suitable for Chinese text tasks was used as the baseline model of our experiment. The experiment is mainly for "Text-level Chinese text classification tasks" are improved, which mainly combines Tapt (Task-Adaptive Pretraining) and "Multi-Sample Dropout method" to improve the model, and compare the experimental results, experimental data sets and model scoring standards Both are consistent with the official WWMBERT model using Accuracy as the scoring standard. The official WWMBERT model uses the maximum and average values of multiple experimental results as the experimental scores. The development set was 97.70% (97.50%) on the "text-level Chinese text classification task". and 97.70% (97.50%) of the test set. After comparing the results of the experiments in this paper, the development set increased by 0.35% (0.5%) and the test set increased by 0.31% (0.48%). The original baseline model has been significantly improved.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.12 no.7
• /
• pp.1243-1248
• /
• 2008

As development in technology of bioinformatics recently mates it possible to operate micro-level experiments, we can observe the expression pattern of total genome through on chip and analyze the interactions of thousands of genes at the same time. In this thesis, we used CDNA microarrays of 3840 genes obtained from neuronal differentiation experiment of cortical stem cells on white mouse with cancer. It analyzed and compared performance of each of the experiment result using existing DT, NB, SVM and multi-perceptron neural network classifier combined the similar scale combination method after constructing class classification model by extracting significant gene list with a similar scale combination method proposed in this paper through normalization. Result classifying in Multi-Perceptron neural network classifier for selected 200 genes using combination of PC(Pearson correlation coefficient) and ED(Euclidean distance coefficient) represented the accuracy of 98.84%, which show that it improve classification performance than case to experiment using other classifier.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.49 no.4
• /
• pp.77-84
• /
• 2012

Face classification has wide applications in intelligent video surveillance, content retrieval, robot vision, and human-machine interface. Pose and expression changes, and arbitrary illumination are typical problems for face recognition. When the face is captured at a distance, the image quality is often degraded by blurring and noise corruption. This paper investigates the efficacy of multi-classifier decision level fusion for face classification based on the photon-counting linear discriminant analysis with two different cost functions: Euclidean distance and negative normalized correlation. Decision level fusion comprises three stages: cost normalization, cost validation, and fusion rules. First, the costs are normalized into the uniform range and then, candidate costs are selected during validation. Three fusion rules are employed: minimum, average, and majority-voting rules. In the experiments, unfocusing and motion blurs are rendered to simulate the effects of the long distance environments. It will be shown that the decision-level fusion scheme provides better results than the single classifier.

• Proceedings of the KSRS Conference
• /
• 2007.03a
• /
• pp.113-119
• /
• 2007

This article deals mainly with two objectives viz, 1) the potentiality of very high-resolution(VHR) multi-spectral and pan chromatic QuickBird satellite data in resources mapping over moderate resolution satellite data (IRS LISS III) and 2) the advantages of using object oriented classification method of eCognition software in land use and land cover analysis over the ISODATA classification method. These VHR data offers widely acceptable metric characteristics for cartographic updating and increase our ability to map land use in geometric detail and improve accuracy of local scale investigations. This study has been carried out in the Sukkalampatti mini-watershed, which is situated in the Eastern Ghats of Tamil Nadu, India. The eCognition object oriented classification method succeeded in most cases to achieve a high percentage of right land cover class assignment and it showed better results than the ISODATA pixel based one, as far as the discrimination of land cover classes and boundary depiction is concerned.

• Korean Journal of Remote Sensing
• /
• v.15 no.1
• /
• pp.39-50
• /
• 1999

It is complicate and time-consuming process to classify a multi-band satellite imagery according to the application. In addition, classification rate sensitively depends on the selection of training data set and features in a supervised classification process. This paper introduced a classification network adopting a fuzzy-based $\gamma$-model in order to select a training data set and to extract feature which highly contribute to an actual classification. The features used in the classification were gray-level histogram, textures, and NDVI(Normalized Difference Vegetation Index) of target imagery. Moreover, in order to minimize the errors in the classification network, the Gradient Descent method was used in the training process for the $\gamma$-parameters at each code used. The trained parameters made it possible to know the connectivity of each node and to delete the void features from all the possible input features.

• International Journal of Knowledge Content Development & Technology
• /
• v.7 no.4
• /
• pp.71-84
• /
• 2017

The paper discusses the nature of Knowledge Organization Systems (KOSs) and shows how these can support digital library users. It demonstrates processes related to integration of KOS like the Dewey Decimal Classification, $22^{nd}$ edition (DDC22) in DSpace software (http://www.dspace.org/) for organizing and retrieving (browsing and searching) scholarly objects. An attempt has been made to use the DDC22 available in Bengali language and highlights the required mechanisms for system-level integration. It may help a repository administrator to build an IDR (Institutional Digital Repository) integrated with SKOS-enabled multilingual subject access systems for supporting subject descriptors based indexing (DC.Subject metadata element), structured navigation (browsing) and efficient searching.

• Asia pacific journal of information systems
• /
• v.1 no.1
• /
• pp.63-86
• /
• 1991

In this paper, our concern is the artificial neural network-based patten classification, when can resolve the difficulties in the Autoregressive Moving Average(ARMA) model identification problem To effectively classify a time series into an approriate ARMA model, we adopt the Multi-layered Backpropagation Network (MLBPN) as a pattern classifier, and Extended Sample Autocorrelation Function (ESACF) as a feature extractor. To improve the classification power of MLBPN's we suggest an integrated neural network system which consists of an AR Network and many small-sized MA Networks. The output of AR Network which will gives the MA order. A step-by-step training strategy is also suggested so that the learned MLBPN's can effectively ESACF patterns contaminated by the high level of noises. The experiment with the artificially generated test data and real world data showed the promising results. Our approach, combined with a statistical parameter estimation method, will provide a way to the automation of ARMA modeling.

• Journal of Korea Multimedia Society
• /
• v.24 no.5
• /
• pp.695-700
• /
• 2021

The Faster RCNN-based shellfish recognition algorithm is introduced for shellfish recognition studies that currently do not have any deep learning-based algorithms in a practical setting. The original feature extraction module is replaced by DenseNet, which fuses multi-level feature data and optimises the NMS algorithm, network depth and merging method; overcoming the omission of shellfish overlap, multiple shellfish and insufficient light, effectively solving the problem of low shellfish classification accuracy. In the complexifier test environment, the test accuracy was improved by nearly 4%. Higher testing accuracy was achieved compared to the original testing algorithm. This provides favourable technical support for future applications of the improved Faster RCNN approach to seafood quality classification.