Text clustering is a hot and essential topic in data mining and information retrieval. This paper proposed a KP-FCM clustering method, which used the key phrases as text features and applied the Fuzzy c-means (FCM) as clustering algorithm. In this method, key phrases were extracted by an algorithm based on suffix array. Experimental results on two standard text clustering benchmark corpuses, OHSUMED (English) and the SOGOU corpus (Chinese) showed that this KP-FCM algorithm outperformed STC-10, Lingo in terms of overall precision, overall recall and overall F-Measure. This indicated that the approach is very effective both in English and Chinese environments. And what's more, since this method was based on key phrases, it could get a readable label of each cluster, which would make the users browse online web search results or volume files more conveniently.
The widespread scale effect always generates significant changes in the properties or characteristics of management objects with different observation scales. Thus, this paper studies the scale transformation mechanism problem of management objects. The observation scale hierarchy (management scale) with clear management objectives could automatically be recognized through changing the observation scales, in order to improve the practical management efficiency. Firstly, an intelligent computing framework based on the scale transformation is established, which reduces the over-dependency of human involvement in traditional scale transformation methods. Then, the scale characteristic reasoning inference is put forward to improve the knowledge acquisition mechanism of scale transformation. Finally, a knowledge acquisition algorithm based on the variable-scale clustering (KAVSC) is proposed. Experiments selected the multiple products inventory data of a manufacturing enterprise from 1 January 2015 to 31 December 2017. The experiment results illustrate that the proposed algorithm KAVSC is able to accurately recognize different management scale levels and scale characteristics of each product, which could effectively support managers making differentiated inventory management plans.
Axles are important part of railway locomotives and vehicles. Periodic ultrasonic inspection of axles can effectively detect and monitor axle fatigue cracks. However, in the axle press-fit zone, the complex interface contact condition reduces the signal-noise ratio (SNR). Therefore, the probability of false positives and false negatives increases. In this work, a novel wavelet threshold function is created to remove noise and suppress press-fit interface echoes in axle ultrasonic defect detection. The exponential threshold function proposed by Andria [1] can't get a gradual curve for later optimum searching process; and the novel wavelet threshold function with two variables is designed to ensure the precision of optimum searching process. Based on the positive correlation between the correlation coefficient and SNR [2] and with the experiment phenomenon (shown in Fig. 2) that the defect and the press-fit interface echo have different axle-circumferential correlation characteristics, a discrete optimum searching process for two undetermined variables in novel wavelet threshold function is conducted. The performance of the proposed method is assessed by comparing it with traditional threshold methods using real data. The statistic results of the amplitude and the peak SNR of defect echoes show that the proposed wavelet threshold denoising method not only maintains the amplitude of defect echoes but also has a higher peak SNR.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.