• Journal of KIISE:Software and Applications
• /
• v.30 no.1_2
• /
• pp.81-92
• /
• 2003

Generalized LR parsing is one of the enhanced LR parsing methods so that it overcome the limit of one-way linear stack of the traditional LR parser using graph-structured stack, and it has been playing an important role of a firm starting point to generate other variations for NL parsing equipped with various mechanisms. In this paper, we propose a conditional Action Model that can solve the problems of conventional probabilistic GLR methods. Previous probabilistic GLR parsers have used relatively limited contextual information for disambiguation due to the high complexity of internal GLR stack. Our proposed model uses Surface Phrasal Types representing the structural characteristics of the parse for its additional contextual information, so that more specified structural preferences can be reflected into the parser. Experimental results show that our GLR parser with the proposed Conditional Action Model outperforms the previous methods by about 6-7% without any lexical information, and our model can utilize the rich stack information for syntactic disambiguation of probabilistic LR parser.

• Communications for Statistical Applications and Methods
• /
• v.27 no.3
• /
• pp.327-347
• /
• 2020

We consider a credit portfolio with highly skewed exposures. In the portfolio, small number of obligors have very high exposures compared to the others. For the Bernoulli mixture model with highly skewed exposures, we propose a new importance sampling scheme to estimate the tail loss probability over a threshold and the corresponding expected shortfall. We stratify the sample space of the default events into two subsets. One consists of the events that the obligors with heavy exposures default simultaneously. We expect that typical tail loss events belong to the set. In our proposed scheme, the tail loss probability and the expected shortfall corresponding to this type of events are estimated by a conditional Monte Carlo, which results in variance reduction. We analyze the properties of the proposed scheme mathematically. In numerical study, the performance of the proposed scheme is compared with an existing importance sampling method.

• Proceedings of the Korean Geotechical Society Conference
• /
• 2003.03a
• /
• pp.123-130
• /
• 2003

The estimation of key soil properties and subsequent quantitative assessment of the associated uncertainties has always been an important issue in geotechnical engineering. It is well recognized that soil properties vary spatially as a result of depositional and post-depositional processes. The stochastic nature of spatially varying soil properties can be treated as a random field. A practical statistical approach that can be used to systematically model various sources of uncertainty is presented in the context of reliability analysis of slope stability Newly developed expressions for probabilistic characterization of soil properties incorporate sampling and measurement errors, as well as spatial variability and its reduced variance due to spatial averaging. Reliability analyses of the probability of slope failure using the different statistical representations of soil properties show that the incorporation of spatial correlation and conditional simulation leads to significantly lower probability of failure than obtained using simple random variable approach.

• Journal of Power System Engineering
• /
• v.13 no.2
• /
• pp.63-70
• /
• 2009

This paper discusses the influence on long-tenn predictions of the ship response in ocean by using the Global Wave Statistics data, GWS, and wave information from the remote sensing satellites. GWS's standard scatter diagrams of significant wave height and zero-crossing wave period are suggested to be corrected to a round number of 0.01/1000 fitted with a statistical analytic model of the conditional lognormal distribution for zero-crossing wave period. The GEOSAT satellite data are utilized which presented by I. R. Young and G. J. Holland (1996, named as GEOSAT data). At first, qualities of this data are investigated, and statistical characteristic trends are studied by means of applying known probability distribution functions. The wave height data of GEOSAT are compared to the data observed onboard merchant ships, the data observed by measure instrument installed on the ocean-going container ship and so on. To execute a long-tenn prediction of ship response, joint probability functions between wave height and wave period are introduced, therefore long-term statistical predictions are executed by using the functions.

• Communications for Statistical Applications and Methods
• /
• v.5 no.3
• /
• pp.905-926
• /
• 1998

We propose a modified mid P-value method to reduce the conservativeness for the inference of conditional associations in three-way contingency tables. This improves the ordinary mfd P-value method. For $2{\times} 2${\times} K$ tables, we propose confidence intervals for an assumed common odds ratio based on inverting two separate one-sided tests using the modified mid P-value. Also, an alternative and usually even better ways of constructing intervals, based on Inverting a two-sided test, are presented. The actual probability of coverage of a 100($1-\alpha$)% confidence interval is centered about the nominal level, but the modified mid P-value approach gives actual coverage probability even closer to the nominal level than the ordinary mid P-value approach.

• Proceedings of the IEEK Conference
• /
• 2003.07c
• /
• pp.2787-2790
• /
• 2003

In this paper, we predict tracking performance of the multiple hypothesis tracking (MHT) algorithm. The MHT algorithm is known to be an optimal Bayesian approach and is superior to asly other tracking filters because it takes into account the events that the measurements can be originated from new targets and false alarms 3s well as interesting targets. In the MHT algorithm, a number of candidate hypotheses are generated and evaluated later as more data are received. The probability of each candidate hypotheses is approximately evaluated by using the hybrid conditional average approach (HYCA). We performed numerical experiments to show the validity of our performance prediction.

• Journal of the Korean Statistical Society
• /
• v.28 no.4
• /
• pp.457-468
• /
• 1999

Autologistic lattice process is used to model binary spatial data. A conditional probability is derived for the incomplete data where the lattice consists of partially yet systematically observed sites. This result, which is interesting in its own right, is in turn applied to area prediction in the plane.

• Journal of the Korean Mathematical Society
• /
• v.55 no.2
• /
• pp.447-469
• /
• 2018

Types (over parameters) in the theory of atomless random variable structures correspond precisely to (conditional) distributions in probability theory. Moreover, the logic (resp. metric) topology on the type space corresponds to the topology of weak (resp. strong) convergence of distributions. In this paper, we study metrics between types. We show that type spaces under $d^{\ast}-metric$ are isometric to Wasserstein spaces. Using optimal transport theory, two formulas for the metrics between types are given. Then, we give a new proof of an integral formula for the Wasserstein distance, and generalize some results in optimal transport theory.

• Communications for Statistical Applications and Methods
• /
• v.5 no.1
• /
• pp.107-120
• /
• 1998

This research aims here to develop a certain extended double sampling plan, EDS, which is an extension of ordinary double sampling plan in the sense that the second-stage sampling effort and second-stage critical value are allowed to depend on the point at which the first-stage continuation region is traversed. For purpose of comparison, single sampling plan, optimal ordinary double sampling plan(ODS) and sequential probability ratio test are considered with the same overall error rates, respectively. It is observed that the EDS idea allows less sampling effort than the optimal ODS.

• Communications of the Korean Mathematical Society
• /
• v.16 no.2
• /
• pp.291-296
• /
• 2001

For randomly indexed sums of the form (Equation. See Full-text), where {X(sub)ni, i$\geq$1, n$\geq$1} are random variables, {N(sub)n, n$\geq$1} are suitable conditional expectations and {b(sub)n, n$\geq$1} are positive constants, we establish a general weak law of large numbers. Our result improves that of Hong [3].