Earticle

본 논문에서는 거리기반 비선형 매니폴드 학습인 ISOMap에 노이즈에 강건한 거리함수를 결합한 새로운 차원축소 방법을 제안한다. ISOMap은 주성분분석이나 선형판별분석과 같은 선형 부분공간 분석법과 는 달리 데이터가 가지 는 비선형 구조를 고려한 그래프 기반 거리함수를 사용함으로써 고차원 입력공간에서 데이 터가 형성하는 비선형의 저차원 매니폴드를 찾아준다. 그러나 기본적인 ISOMap은 클래스 정보를 사용하지 않는 비 교사 학습을 수행할 뿐 아니라, 새로운 데이터에 대한 매핑함수를 제공해 주지 않으므로 패턴인식에 쉽게 적용되기 어렵고 학습 데이터가 충분히 조밀하지 않은 경우에는 노이즈에 강건하지 못한 문제점을 가지고 있다. 본 논문에 서는 이러한 문제점을 해결하기 위하여 클래스 정보를 활용한 교사학습을 수행하면서, 클래스 내 데이터의 변형 정 보를 활용하여 노이즈에 강건한 거리함수를 새롭게 정의함으로써 패턴인식에 효과적인 새로운 비선형 매니폴드 학습법을 제안한다. 제안하 는 방법을 노이즈를 추가한 벤치마크 데이터에 적용하여 그 성능을 확인하였다.

In this paper, we propose a novel dimension reduction method by modifying ISOMap with a noise-tolerant distance function. The ISOMap is the well known distance-based nonlinear manifold learning method. Unlike the linear subspace analysis methods such as PCA and LDA, it can find a low dimensional manifold that can represent nonlinear structure of data set by using graph-based distance function. However, the original ISOMap is not appropriate for patten classification problems because it does unsupervised learning without classification information and is likely to be sensitive to noises when the data set is not sufficiently dense. In order to solve the problems, we propose a new supervised manifold learning method, which uses a new distance function that is robust to input noise. The performance of the proposed method is confirmed by the experiments using the benchmark data with additional noises.

Modulation type is one of the most important characteristics used in radar signal recognition. This paper proposes a method to realize modulation identification. This algorithm applies wavelet transformation to the signal, and then uses manifold learning method to reduce the high dimension and extracts the recognition feature. The proper threshold value is set as the classifier to achieve the purpose of recognizing 5 kinds of signals （2ASK, 2FSK, 2PSK, LFM, CP）in Gauss white noise environment. The algorithm requires priori signal information no other than signal-to-noise rate. Simulation result indicates the algorithm achieves good performance.

Modulation type is one of the most important characteristics used in signal recognition. An algorithm to realize signal modulation identification is proposed in this paper. We applied wavelet transformation and STFT to the signal, and then used manifold learning method to reduce the high dimension and extracted the recognition feature. The proper threshold value was set as the classifier to achieve the purpose of recognizing 4 kinds of signals (MASK, MFSK, MPSK,QAM) in Gauss white noise environment. The algorithm requires priori signal information no other than signal-to-noise rate. Simulation result indicates the algorithm achieves good performance.

Wind power prediction has become a hotspot in recent years. The parameters relevant to the wind power are considerable and complexity. Dimension reduction has become another hotspot. The traditional methods utilize linear methods to reduce the dimension of measured data. However, data that located in high-dimensional space often have nonlinear structure. So, we consider the original data present manifold structures and introduce manifold learning methods to extract the important information. In this paper, we utilize LLE algorithm and Elman neural network to establish the wind power prediction model. The experiment results demonstrate the excellence of our method. Finally, we chose different algorithm parameters to complete the experiments and got the roughly optimal parameters. In addition, our method can be applied to similar fields.