Papers by Keyword: Cosine Similarity

Abstract: Two new algorithms for nonlinear damage detection are proposed based on linear model with autoregressive moving average (ARMA) in this paper. Firstly, a novel DSF is defined and the DSFs are identified and classified followed by cluster analysis or Bayesian discrimination. Secondly, the performances of the presented algorithms are evaluated and verified by the experimental data of a three-story building structure. Finally, the illustrated results show the algorithms are efficient tools for nonlinear damage detection. They grant a higher accuracy and improve the reliability of nonlinear damage detection whilst reducing computational costs. It can thus be inferred that the proposed algorithms are applicable for Structural Health Monitoring (SHM) in situ.

Abstract: A novel method of vehicle type recognition based on template matching is proposed to improve the real-time performance of the vehicle type recognition in real traffic scenes. GRM is applied and the template is normalized for realizing parallel template matching. Then, we realize the rapid vehicle type recognition through lookup tables by the hierarchical index of vehicle type template with k-means clustering and size normalization processing. The results show that the algorithm can recognize vehicle type in traffic scenes efficiently.

Abstract: Since millions of documents are available on the Internet, some documents contain similar content but they are written in different languages by various authors. Unfortunately, the existing search engines do not support to all documents that are relevant to a single language query. Therefore, several researchers have put a huge effort to overcome such a problem. The major problems of a cross language search engine include 1) how to store information in a unify model and represent information of multiple languages documents effectively and 2) how to rank the retrieved multiple language documents and present to a user in the right order. This paper overcomes the first problem using an ontology model and we present a new ranking technique for a cross language information retrieval system (CLIR). Keyword weighting scheme in an ontology and document sections are introduced. Cosine similarity formula is modified to particularly support CLIR. The experimental results show the modified formula obtains more efficient ranking results than the existing method.

Abstract: In case of mechanical system health monitoring, a need to develop normal-knowledge based novelty detection techniques is increasing. The negative selection algorithm, which is inspired from the operation mechanism of human immune system, is one of such approaches. Our approach is to apply the idea for the anomaly detection in the vibration time series of the rotor system. A real-valued negative selection algorithm based on Euclidean distance, as well as cosine similarity, has been implemented. By means of adding the corresponding coverage radius to each antibody elements, the detection efficiency of each antibody element is increased. The detection efficiency is evaluated with simulated data as well as vibration signal sampled from one rotor system. The results indicate that the algorithm can efficiently detect the anomaly in time series data. Moreover, the number of detectors in antibody set is less enough for potential application in online signal monitoring.