Study on Feature Dimension Reduction Method of Emergency Topic Model Based on Improved CHI and LSA

Liang Meiyu,Du Junping,Jia Yingmin,Sun Zengqi 제어로봇시스템학회 2011 제어로봇시스템학회 국제학술대회 논문집 Vol.2011 No.10

According to some flaws in the existing feature dimension reduction methods, a new method of two-step combined feature dimension reduction based on improved CHI algorithm and LSA algorithm is proposed in this paper. First, apply the improved CHI algorithm to realize the initial feature selection, resolve the problem of high dimension and sparseness in the feature space to a certain extent, and then use the LSA algorithm to extract the semantic structures in the initial feature space, and map it into the semantic feature space and realize the second dimension reduction. Experimental results indicate that this method of feature dimension reduction has a better performance, further improving the effect of topic tracking.

Novel Intent based Dimension Reduction and Visual Features Semi-Supervised Learning for Automatic Visual Media Retrieval

kunisetti, Subramanyam,Ravichandran, Suban International Journal of Computer ScienceNetwork S 2022 International journal of computer science and netw Vol.22 No.6

Sharing of online videos via internet is an emerging and important concept in different types of applications like surveillance and video mobile search in different web related applications. So there is need to manage personalized web video retrieval system necessary to explore relevant videos and it helps to peoples who are searching for efficient video relates to specific big data content. To evaluate this process, attributes/features with reduction of dimensionality are computed from videos to explore discriminative aspects of scene in video based on shape, histogram, and texture, annotation of object, co-ordination, color and contour data. Dimensionality reduction is mainly depends on extraction of feature and selection of feature in multi labeled data retrieval from multimedia related data. Many of the researchers are implemented different techniques/approaches to reduce dimensionality based on visual features of video data. But all the techniques have disadvantages and advantages in reduction of dimensionality with advanced features in video retrieval. In this research, we present a Novel Intent based Dimension Reduction Semi-Supervised Learning Approach (NIDRSLA) that examine the reduction of dimensionality with explore exact and fast video retrieval based on different visual features. For dimensionality reduction, NIDRSLA learns the matrix of projection by increasing the dependence between enlarged data and projected space features. Proposed approach also addressed the aforementioned issue (i.e. Segmentation of video with frame selection using low level features and high level features) with efficient object annotation for video representation. Experiments performed on synthetic data set, it demonstrate the efficiency of proposed approach with traditional state-of-the-art video retrieval methodologies.

점진적 모델에 기반한 다채널 시계열 데이터 EEG의 특징 분석

김선희,양형정,Kan Seww Ng,정종문 한국정보처리학회 2009 정보처리학회논문지. 소프트웨어 및 데이터 공학 Vol.16 No.1

BCI technology is to control communication systems or machines by brain signal among biological signals followed by signal processing. For the implementation of BCI systems, it is required that the characteristics of brain signal are learned and analyzed in real-time and the learned characteristics are applied. In this paper, we detect feature vector of EEG signal on left and right hand movements based on incremental approach and dimension reduction using the detected feature vector. In addition, we show that the reduced dimension can improve the classification performance by removing unnecessary features. The processed data including sufficient features of input data can reduce the time of processing and boost performance of classification by removing unwanted features. Our experiments using K-NN classifier show the proposed approach 5% outperforms the PCA based dimension reduction. BCI 기술은 생체신호인 뇌파를 수집하여 신호처리를 거친 후 실질적인 기기제어 및 통신 시스템 등을 제어하는 시스템 관련 기술이다. BCI 시스템 구현을 위해서는 뇌파의 특성을 실시간으로 분석하여 학습 시키고 학습된 뇌파의 특성을 적용하는 단계가 요구된다. 본 논문에서는 EEG 데이터를 효율적으로 분석하기 위해 점진적으로 갱신되는 주성분 분석을 이용하여 왼손/오른손 동작에 영향을 미치는 EEG 신호의 특징을 찾고, 이를 반영하여 데이터의 차원을 축소한다. 입력 자료의 특징을 충분히 포함하면서 낮은 차원을 가지는 데이터를 이용한다면 분류를 위한 계산량을 감소시킬 수 있을 뿐만 아니라 불필요한 특징을 제거함으로써 분류 성능을 향상 시킬 수 있다. 본 논문에서는 점진적으로 갱신되는 주성분 분석을 이용하여 데이터의 차원을 축소하고 이에 대한 효율성을 검증하기 위해 K-NN분류기를 이용하여 분류 정확도 측정을 수행하였다. 그 결과 주성분 분석을 이용하여 특징을 추출하고 분류율을 측정한 경우보다 평균 5% 높은 분류 정확율을 보였다.

Wrapper Based Wavelet Feature Optimization for EEG Signals

Girisha Garg,A.P. Mittal,Vijander Singh,J.R.P Gupta 대한의용생체공학회 2012 Biomedical Engineering Letters (BMEL) Vol.2 No.1

Purpose In this paper a computationally efficient wrapper based Wavelet Feature Optimization (WFO) is developed. The algorithm is developed for the classification of high dimensional EEG signals which may suffer from the curse of dimensionality and sub optimal feature selection. Methods The key design phases of the algorithm involve: 1)Feature Transformation of the original EEG signals using Discrete Wavelet Transform; 2) Feature Extraction using the concept of Relative Wavelet Energy (RWE) 3) Selecting the optimal subset of the RWE features using wrapper approach. In contrast to the methods guided by the filter technique of feature selection, this approach uses the wrapper based method to select the optimal and a very low dimensional feature space from the wavelet features. Results The highlight of the algorithm is that in addition to increase the computational efficiency, it also enhances the predictive power of the system without any loss of relevant information. This paper includes the experimentation performed on EEG datasets using WFO algorithm. Conclusions The algorithm gives consistent and excellent performance for the EEG datasets. The feature sets obtained with the help of WFO are also tested using mutual information methods to confirm the optimality of the wavelet feature subset.

Action Recognition with deep network features and dimension reduction

( Lijun Li ),( Shuling Dai ) 한국인터넷정보학회 2019 KSII Transactions on Internet and Information Syst Vol.13 No.2

Action recognition has been studied in computer vision field for years. We present an effective approach to recognize actions using a dimension reduction method, which is applied as a crucial step to reduce the dimensionality of feature descriptors after extracting features. We propose to use sparse matrix and randomized kd-tree to modify it and then propose modified Local Fisher Discriminant Analysis (mLFDA) method which greatly reduces the required memory and accelerate the standard Local Fisher Discriminant Analysis. For feature encoding, we propose a useful encoding method called mix encoding which combines Fisher vector encoding and locality-constrained linear coding to get the final video representations. In order to add more meaningful features to the process of action recognition, the convolutional neural network is utilized and combined with mix encoding to produce the deep network feature. Experimental results show that our algorithm is a competitive method on KTH dataset, HMDB51 dataset and UCF101 dataset when combining all these methods.

Incremental feature extraction based on decision boundaries

Woo, Seongyoun,Lee, Chulhee Elsevier 2018 Pattern recognition Vol.77 No.-

<P><B>Abstract</B></P> <P>Feature extraction is a key algorithm to solve the dimensionality problem. Most feature extraction algorithms use a batch mode, which requires all data available at the same time to calculate new features. Recently, with more available data and advancements of transmission technology, the need for incremental algorithms has increased. In this paper, we propose a gradient descent DBFE method (GDDBFE) that shows a substantial improvement in processing time. Based on this GDDBFE, we then propose an incremental gradient descent decision boundary feature extraction method (IGDDBFE). The proposed IGDDBFE method consists of two steps: updating the decision boundaries and adding discriminately informative features with newly added samples and then updating the feature vectors by incremental eigenvector updates. Experiments with real-world databases show that the proposed method shows improved performance compared to some existing methods.</P> <P><B>Highlights</B></P> <P> <UL> <LI> We developed a gradient based decision boundary feature extraction algorithm for neural networks and its incremental version. </LI> <LI> The proposed method updates the decision boundaries from sequentially added samples and obtains discriminately informative vectors based on the updated decision boundaries. </LI> <LI> When applied to real world databases, it showed noticeably better classification performance than some existing incremental algorithms. </LI> </UL> </P>

차원 축소 알고리즘을 이용한 한국어 자음간 거리 매핑 연구

장선아,오승하,최승진,김민제,박선호 한국청각언어재활학회 2012 Audiology and Speech Research Vol.8 No.1

There have been increasing needs to understand speech signals to reveal auditory processing mechanism. We investigated an optimal dimension reduction algorithm and features for reflecting distance between Korean consonants. First, we reduced features of consonant articulatory phonetics from 15 to 5-6 using three dimension reduction algorithms. This revealed 90% of the distance distribution of Korean consonants. Next, we applied the Isomap method to show the most similar acoustic features of speech banana. Results showed a useful mapping of consonant articulatory distances. For example, we found that consonant /ㅂ(p)/ and /ㄷ(t)/ share 13 out of 15 features of articulatory phonetics, with only the features of bilabial and coronal being different, suggesting that they are the most difficult to discriminate from each other and reflected in the shortest distance between these two consonants in the Isomap method. However, the Isomap method was not effective in distinguishing the acoustic features of fortis from aspirated with limited numbers of acoustic features which are using for English consonants. This study shows that the use of reduction algorithms and the Isomap method can be effective in consonant mapping of articulatory features. These findings suggest that they are useful tools in phonetics research when combined with acoustic phonetics data and real voice data.

Hyperspectral Image Classification based on Co-training

Zhijun Zheng,Yanbin Peng 보안공학연구지원센터 2015 International Journal of Hybrid Information Techno Vol.8 No.12

The abundant information available in hyperspectral image has provided important opportunities for land-cover classification and recognition. However, “Curse of dimensionality” and small training sample set are two difficulties which hinder the improvement of computational efficiency and classification precision. In this paper, we present a co-training based method on hyperspectral image classification. Firstly, two views of samples are generated through two kinds of dimensionality reduction methods. After that, the co-training process is viewed as combinative label propagation over two independent views. Experimental results on real hyperspectral image show that the proposed method has better performance than the other state-of-the-art methods.

Support Vector Machine Based Classification of 3-Dimensional Protein Physicochemical Environments for Automated Function Annotation

Min, Hye-Young,Yu, Seung-Hak,Lee, Tae-Hoon,Yoon, Sung-Roh 대한약학회 2010 Archives of Pharmacal Research Vol.33 No.9

The knowledge of protein functions as well as structures is critical for drug discovery and development. The FEATURE system developed at Stanford is an effective tool for characterizing and classifying local environments in proteins. FEATURE utilizes vectors of a fixed dimension to represent the physicochemical properties around a residue. Functional sites and non-sites are identified by classifying such vectors using the Na$\"{\i}$ve Bayes classifier. In this paper, we improve the FEATURE framework in several ways so that it can be more flexible, robust and accurate. The new tool can handle vectors of a user-specified dimension and can suppress noise effectively, with little loss of important signals, by employing dimensionality reduction. Furthermore, our approach utilizes the support vector machine for a more accurate classification. According to the results of our thorough experiments, the proposed new approach outperformed the original tool by 20.13% and 13.42% with respect to true and false positive rates, respectively.

The Impact of Feature Reduction Techniques on Arabic Document Classification

Abdullah Ayedh,Guanzheng Tan,Hamdi Rajeh 보안공학연구지원센터 2016 International Journal of Database Theory and Appli Vol.9 No.6

Feature reduction are common techniques that used to improve the efficiency and accuracy of the document classification systems. The problems associated with these techniques are the highly dimensionality of the feature space and The difficulty of selecting the important features for understanding the document in question. The document usually consists of several parts and the important features that more closely associated with the topic of the document are appearing in the first parts or repeated in several parts of the document. Therefore, the position of the first appearance of a word and the compactness of the word considered as factors that determine the important features using the information within a document. This study, explored the impact of combining three feature weighting methods that depend on inverse document frequency (IDF), namely, Term frequency (TFiDF), the position of the first appearance of a word (FAiDF), and the compactness of the word (CPiDF) on the classification accuracy. In addition, we have investigated different feature selection techniques, namely, Information gain (IG), Goh and Low (NGL) coefficients, Chi-square Testing (CHI), and Galavotti-Sebastiani-Simi Coefficient (GSS) in order to improve the performance for Arabic document classification system. Experimental analysis on Arabic datasets reveals that the proposed methods have a significant impact on the classification accuracy, and in most cases the FAiDF feature weighting performed better than CPiDF and TFiDF. The results also clearly showed the superiority of the GSS over the other feature selection techniques and achieved 98.39% micro-F1 value when using a combination of TFiDF, FAiDF, and CPiDF as feature weighting method.