• Title/Summary/Keyword: Log likelihood ratio

Search Result 141, Processing Time 0.023 seconds

A Study on Utterance Verification Using Accumulation of Negative Log-likelihood Ratio (음의 유사도 비율 누적 방법을 이용한 발화검증 연구)

  • 한명희;이호준;김순협
    • The Journal of the Acoustical Society of Korea
    • /
    • v.22 no.3
    • /
    • pp.194-201
    • /
    • 2003
  • In speech recognition, confidence measuring is to decide whether it can be accepted as the recognized results or not. The confidence is measured by integrating frames into phone and word level. In case of word recognition, the confidence measuring verifies the results of recognition and Out-Of-Vocabulary (OOV). Therefore, the post-processing could improve the performance of recognizer without accepting it as a recognition error. In this paper, we measure the confidence modifying log likelihood ratio (LLR) which was the previous confidence measuring. It accumulates only those which the log likelihood ratio is negative when integrating the confidence to phone level from frame level. When comparing the verification performance for the results of word recognizer with the previous method, the FAR (False Acceptance Ratio) is decreased about 3.49% for the OOV and 15.25% for the recognition error when CAR (Correct Acceptance Ratio) is about 90%.

Joint Performance of Demodulation and Decoding with Regard to Log-Likelihood Ratio Approximation (대수우도비 근사화에 따른 복조와 복호의 결합 성능)

  • Park, Sung-Joon;Jo, Myung-Suk
    • The Journal of Korean Institute of Communications and Information Sciences
    • /
    • v.41 no.12
    • /
    • pp.1736-1738
    • /
    • 2016
  • In modern digital communication systems adapting high-order modulation and high performance channel code, log-likelihood ratios involving the repeated calculations of the logarithm of sum of exponential functions are necessary for demodulation and decoding. In this paper, the approximation methods called Min and MinC are applied to demodulation and decoding together and their complexity and joint performance are analyzed.

Performance Analysis Based On Log-Likelihood Ratio in Orthogonal Code Hopping Multiplexing Systems Using Multiple Antennas (다중 안테나를 사용한 직교 부호 도약 다중화 시스템에서 로그 우도비 기반 성능 분석)

  • Jung, Bang-Chul;Sung, Kil-Young;Shin, Won-Yong
    • Journal of the Korea Institute of Information and Communication Engineering
    • /
    • v.15 no.12
    • /
    • pp.2534-2542
    • /
    • 2011
  • In this paper, we show that performance can be improved by using multiple antennas in the conventional orthogonal code hopping multiplexing (OCHM) scheme, which was proposed for accommodating a larger number of users with low channel activities than the number of orthogonal codewords used in code division multiple access (CDMA)-based communication systems through downlink statistical multiplexing. First, we introduce two different types of OCHM systems together with orthogonal codeword allocation strategies, and then derive their mathematical expression for log-likelihood ratio (LLR) values according to the two different schemes. Next, when a turbo encoder based on the LLR computation is used, we evaluate performance on the frame error rate (FER) for the aformentioned OCHM system. For comparison, we also show performance for the existing symbol mapping method using multiple antennas, which was used in 3GPP standards. As a result, it is shown that our OCHM system with multiple antennas based on the proposed orthogonal codeword allocation strategy leads to performance gain over the conventional system---energy required to satisfy a target FER is significantly reduced.

Safety Performance Functions for Central Business Districts Using a Zero-Inflated Model (영과잉을 고려한 중심상업지구 교통사고모형 개발에 관한 연구)

  • Lee, Sang Hyuk;Woo, Yong Han
    • International Journal of Highway Engineering
    • /
    • v.18 no.4
    • /
    • pp.83-92
    • /
    • 2016
  • PURPOSES : The purpose of this study was to develop safety performance functions (SPFs) that use zero-inflated negative binomial regression models for urban intersections in central business districts (CBDs), and to compare the statistical significance of developed models against that of regular negative binomial regression models. METHODS : To develop and analyze the SPFs of intersections in CBDs, data acquisition was conducted for dependent and independent variables in areas of study. We analyzed the SPFs using zero-inflated negative binomial regression model as well as regular negative binomial regression model. We then compared the results by analyzing the statistical significance of the models. RESULTS : SPFs were estimated for all accidents and injury accidents at intersections in CBDs in terms of variables such as AADT, Number of Lanes at Major Roads, Median Barriers, Right Turn with an Exclusive Turn Lane, Turning Guideline, and Front Signal. We also estimated the log-likelihood at convergence and the likelihood ratio of SPFs for comparing the zero-inflated model with the regular model. In he SPFs, estimated log-likelihood at convergence and the likelihood ratio of the zero-inflated model were at -836.736, 0.193 and -836.415, 0.195. Also estimated the log-likelihood at convergence and likelihood ratio of the regular model were at -843.547, 0.187 and -842.631, 0.189, respectively. These figures demonstrate that zero-inflated negative binomial regression models can better explain traffic accidents at intersections in CBDs. CONCLUSIONS : SPFs that use a zero-inflated negative binomial regression model demonstrate better statistical significance compared with those that use a regular negative binomial regression model.

Butterfly Log-MAP Decoding Algorithm

  • Hou, Jia;Lee, Moon Ho;Kim, Chang Joo
    • Journal of Communications and Networks
    • /
    • v.6 no.3
    • /
    • pp.209-215
    • /
    • 2004
  • In this paper, a butterfly Log-MAP decoding algorithm for turbo code is proposed. Different from the conventional turbo decoder, we derived a generalized formula to calculate the log-likelihood ratio (LLR) and drew a modified butterfly states diagram in 8-states systematic turbo coded system. By comparing the complexity of conventional implementations, the proposed algorithm can efficiently reduce both the computations and work units without bit error ratio (BER) performance degradation.

Utterance Verification using Phone-Level Log-Likelihood Ratio Patterns in Word Spotting Systems (핵심어 인식기에서 단어의 음소레벨 로그 우도 비율의 패턴을 이용한 발화검증 방법)

  • Kim, Chong-Hyon;Kwon, Suk-Bong;Kim, Hoi-Rin
    • Phonetics and Speech Sciences
    • /
    • v.1 no.1
    • /
    • pp.55-62
    • /
    • 2009
  • This paper proposes an improved method to verify a keyword segment that results from a word spotting system. First a baseline word spotting system is implemented. In order to improve performance of the word spotting systems, we use a two-pass structure which consists of a word spotting system and an utterance verification system. Using the basic likelihood ratio test (LRT) based utterance verification system to verify the keywords, there have been certain problems which lead to performance degradation. So, we propose a method which uses phone-level log-likelihood ratios (PLLR) patterns in computing confidence measures for each keyword. The proposed method generates weights according to the PLLR patterns and assigns different weights to each phone in the process of generating confidence measures for the keywords. This proposed method has shown to be more appropriate to word spotting systems and we can achieve improvement in final word spotting accuracy.

  • PDF

General Log-Likelihood Ratio Expression and Its Implementation Algorithm for Gray-Coded QAM Signals

  • Kim, Ki-Seol;Hyun, Kwang-Min;Yu, Chang-Wahn;Park, Youn-Ok;Yoon, Dong-Weon;Park, Sang-Kyu
    • ETRI Journal
    • /
    • v.28 no.3
    • /
    • pp.291-300
    • /
    • 2006
  • A simple and general bit log-likelihood ratio (LLR) expression is provided for Gray-coded rectangular quadrature amplitude modulation (R-QAM) signals. The characteristics of Gray code mapping such as symmetries and repeated formats of the bit assignment in a symbol among bit groups are applied effectively for the simplification of the LLR expression. In order to reduce the complexity of the max-log-MAP algorithm for LLR calculation, we replace the mathematical max or min function of the conventional LLR expression with simple arithmetic functions. In addition, we propose an implementation algorithm of this expression. Because the proposed expression is very simple and constructive with some parameters reflecting the characteristic of the Gray code mapping result, it can easily be implemented, providing an efficient symbol de-mapping structure for various wireless applications.

  • PDF

Mutual Information and Redundancy for Categorical Data

  • Hong, Chong-Sun;Kim, Beom-Jun
    • Communications for Statistical Applications and Methods
    • /
    • v.13 no.2
    • /
    • pp.297-307
    • /
    • 2006
  • Most methods for describing the relationship among random variables require specific probability distributions and some assumptions of random variables. The mutual information based on the entropy to measure the dependency among random variables does not need any specific assumptions. And the redundancy which is a analogous version of the mutual information was also proposed. In this paper, the redundancy and mutual information are explored to multi-dimensional categorical data. It is found that the redundancy for categorical data could be expressed as the function of the generalized likelihood ratio statistic under several kinds of independent log-linear models, so that the redundancy could also be used to analyze contingency tables. Whereas the generalized likelihood ratio statistic to test the goodness-of-fit of the log-linear models is sensitive to the sample size, the redundancy for categorical data does not depend on sample size but its cell probabilities itself.

Low-Complexity Soft-MIMO Detection Algorithm Based on Ordered Parallel Tree-Search Using Efficient Node Insertion (효율적인 노드 삽입을 이용한 순서화된 병렬 트리-탐색 기반 저복잡도 연판정 다중 안테나 검출 알고리즘)

  • Kim, Kilhwan;Park, Jangyong;Kim, Jaeseok
    • The Journal of Korean Institute of Communications and Information Sciences
    • /
    • v.37A no.10
    • /
    • pp.841-849
    • /
    • 2012
  • This paper proposes an low-complexity soft-output multiple-input multiple-output (soft-MIMO) detection algorithm for achieving soft-output maximum-likelihood (soft-ML) performance under max-log approximation. The proposed algorithm is based on a parallel tree-search (PTS) applying a channel ordering by a sorted-QR decomposition (SQRD) with altered sort order. The empty-set problem that can occur in calculation of log-likelihood ratio (LLR) for each bit is solved by inserting additional nodes at each search level. Since only the closest node is inserted among nodes with opposite bit value to a selected node, the proposed node insertion scheme is very efficient in the perspective of computational complexity. The computational complexity of the proposed algorithm is approximately 37-74% of that of existing algorithms, and from simulation results for a $4{\times}4$ system, the proposed algorithm shows a performance degradation of less than 0.1dB.