Non-negative matrix factorization(NMF)is a technique for dimensionality reduction by placing non-negativity constraints onthe matrix.Based on the PARAFAC model,NMF was extended for three-dimension data decomposition.The three-dimension non-negative matrix factorization(NMF3)algorithm,which was concise and easy to implement,was given in this paper.The NMF3algorithm implementation was based on elements but not on vectors.It could decompose a data array directly without unfolding,which was not similar to that the traditional algorithms do.It has been applied to the simulated data array decomposition andobtained reasonable results.It showed that NMF3 could be introduced for curve resolution in chemometrics. 相似文献
JavaStatSoft, implemented using Java, is an open-source software for performing statistics. The current features of JavaStatSoft
are described. Several well-known patterns the software used are introduced. In addition, based on these patterns and the
experience in developing the software, a new framework, referred to as Data Analysis Module, is proposed to help programmers
develop a user-friendly software for data analysis. 相似文献
Summary This paper deals with structural damage detection using measured frequency response functions (FRF) as input data to artificial
neural networks (ANN). A major obstacle, the impracticality of using full-size FRF data with ANNs, was circumvented by applying
a data-reduction technique based on principal component analysis (PCA). The compressed FRFs, represented by their projection
onto the most significant principal components, were used as the ANN input variables instead of the raw FRF data. The output
is a prediction of the actual state of the specimen, i.e. healthy or damaged. A further advantage of this particular approach
is its ability to deal with relatively high measurement noise, which is a common occurrence when dealing with industrial structures.
The methodology was applied to detect three different states of a space antenna: reference, slight mass damage and slight
stiffness damage. About 600 FRF measurements, each with 1024 spectral points, were included in the analysis. Six 2-hidden
layer networks, each with an individually-optimised architecture for a specific FRF reduction level, were used for damage
detection. The results showed that it was possible to distinguish between the three states of the antenna with good accuracy,
subject to using an adequate number of principal components together with a suitable neural network configuration. It was
also found that the quality of the raw FRF data remained a major consideration, though the method was able to filter out some
of the measurement noise. The convergence and detection properties of the networks were improved significantly by removing
those FRFs associated with measurement errors.
Received 9 March 2000; accepted for publication 12 December 2000 相似文献
Given a sample of binary random vectors with i.i.d. Bernoulli(p) components, that is equal to 1 (resp. 0) with probability p (resp. 1−p), we first establish a formula for the mean of the size of the random Galois lattice built from this sample, and a more complex one for its variance. Then, noticing that closed α-frequent itemsets are in bijection with closed α-winning coalitions, we establish similar formulas for the mean and the variance of the number of closed α-frequent itemsets. This can be interesting for the study of the complexity of some data mining problems such as association rule mining, sequential pattern mining and classification. 相似文献
Widely publicized reports of fresh MBAs getting multiple job offers with six-figure annual salaries leave a long-lasting general impression about the high quality of selected business schools. While such spectacular achievement in job placement rightly deserves recognition, one should not lose sight of the resources expended in order to accomplish this result. In this study, we employ a measure of Pareto-Koopmans global efficiency to evaluate the efficiency levels of the MBA programs in Business Week’s top-rated list. We compute input- and output-oriented radial and non-radial efficiency measures for comparison. Among three tier groups, the schools from a higher tier group on average are more efficient than those from lower tiers, although variations in efficiency levels do occur within the same tier, which exist over different measures of efficiency. 相似文献
One of the typical issues in financial literature is that the market tends to be overly pessimistic about value stocks, many of which are past losers. Therefore, over-reactions might capture by measuring earnings surprise vary with past return levels. In this paper, we propose a new index for an effective investment strategy to capture the return-reversal effect using both Data Envelopment Analysis (DEA) and Inverted DEA in order to consider the above characteristics of the market. Our investment strategy using the new index exhibits better performance than the naive return-reversal strategy that only uses past returns or earnings surprise. In addition, the correlations between our new index and commonly used value indices are insignificant, and the value indices cannot represent the over-valued (under-valued) situations perfectly. Hence, considering both proposed and value indices like book-to-price one, we could select value stocks more effectively than by using only one of these indices. 相似文献
We report on ideas, problems and results, which occupied us during the past decade and which seem to extend the frontiers of information theory in several directions. The main contributions concern information transfer by channels. There are also new questions and some answers in new models of source coding. While many of our investigations are in an explorative state, there are also hard cores of mathematical theories. In particular we present a unified theory of information transfer, which naturally incorporates Shannon's theory of information transmission and the theory of identification in the presence of noise as extremal cases. It provides several novel coding theorems. On the source coding side we introduce data compression for identification. Finally we are led beyond information theory to new concepts of solutions for probabilistic algorithms.
The original paper [R. Ahlswede, General theory of information transfer, Preprint 97-118, SFB 343 Diskrete Strukturen in der Mathematik, Universität Bielefeld, 1997] gave to and received from the ZIF-project essential stimulations which resulted in contributions added as GTIT-Supplements “Search and channels with feedback” and “Noiseless coding for multiple purposes: a combinatorial model”.
Other contributions—also to areas initiated—are published in the recent book [R. Ahlswede et al. (Eds.), General Theory of Information Transfer and Combinatorics, Lecture Notes in Computer Science, vol. 4123, Springer, Berlin, 2006].
The readers are advised to study always the pioneering papers in a field—in this case the papers [R. Ahlswede, G. Dueck, Identification via channels, IEEE Trans. Inform. Theory 35 (1989) 15–29; R. Ahlswede, G. Dueck, Identification in the presence of feedback—a discovery of new capacity formulas, IEEE Trans. Inform. Theory 35 (1989) 30–39] on identification. It is not only the most rewarding way to come to new ideas, but it also helps to more quickly grasp the more advanced formalisms without going through too many technicalities. Perhaps also the recent Shannon Lecture [R. Ahlswede, Towards a General Theory of Information Transfer, Shannon Lecture at ISIT in Seattle 13th July 2006, IEEE Information Theory Society Newsletter, 2007], aiming at an even wider scope, gives further impetus. 相似文献