Combining experts for anchorperson shot detection in news videos

Automatic classification of shots extracted by news videos plays an important role in the context of news video segmentation, which is an essential step towards effective indexing of broadcasters digital databases. In spite of the efforts reported by the researchers involved in this field, no techniques providing fully satisfactory performance have been presented until now. In this paper, we propose a multi-expert approach for unsupervised shot classification. The proposed multi-expert system (MES) combines three algorithms that are model-free and do not require a specific training phase. In order to assess the performance of the MES, we built up a database significantly wider than those typically used in the field. Experimental results demonstrate the effectiveness of the proposed approach both in terms of shot classification and of news story detection capability.     

"Development of Time, Speed, and Distance Concepts": Correction to Siegler and Richards.

Reports an error in the original article by Robert S. Siegler and D. Dean Richards (Developmental Psychology, 1979, Vol. 15[3], pp. 288-298). The column headings of Table 1 on p. 291 are incorrect. Each pair of lines in the column headings represents the starting points, stopping points, and total distance traveled by the trains. The corrected table appears below. (The following abstract of this article originally appeared in record 1979-28311-001.) Used the rule-assessment approach to examine understanding of the concepts of time, speed, and distance in 36 5-, 8-, and 11-yr-olds and 12 undergraduates. Parallel tasks were developed for the 3 concepts that allowed specification of whether Ss were relying on time, speed, distance, end point, end time, beginning point, or beginning time cues in making their judgments. It was found that 5-yr-olds understood all 3 concepts in the same way: Whichever train ended farther ahead on the tracks was said to have traveled for the longer time, at the faster speed, and for the greater distance. Undergraduates, at the other extreme, understood all 3 concepts as distinct and separate ideas. The transitional period was marked by specific confusions among the 3 concepts: Time was regularly confused with distance, distance was confused with time, and speed was confused with distance and to some extent with end point. Both speed and distance concepts appeared to be mastered well before the concept of time. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Reports for 1982 of the Regional Associations.

Reports for 1982 are listed for the following regional association meetings: Sixty-Second Annual Meeting of the Western Psychological Association; Fifty-Second Annual Meeting of the Rocky Mountain Psychological Association; Fifty-Fourth Annual Meeting of the Midwestern Psychological Association; Twenty-Eighth Annual Meeting of the Southeastern Psychological Association; Twenty-Eighth Annual Meeting of the Southwestern Psychological Association; Twenty-Second Annual Meeting of the New England Psychological Association; Fifty-Third Annual Meeting of the Eastern Psychological Association; and Seventy-Fourth Annual Meeting of the Southern Society for Philosophy and Psychology. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

条件事件代数在专家系统中的应用研究

条件事件代数是在确保规则概率与条件概率相容的前提下,把布尔代数上的逻辑运算推广到条件事件（规则）集合中的逻辑代数系统,本文介绍了条件事件代数的基本原理和性质,并利用条件事件代数解决了专家系统的规则循环问题,克服了传统谓词逻辑在推理过程中的局限性。     

一种基于粗集理论的动态近似规则挖掘推理方法

提出一种基于粗集理论的,把属性的重要性和属性值的出现频率综合起来进行规则推理的方法.分析了“激活一个,否则离开”原则的优缺点,指出在近似推理中,大前提中的规则数量应该可变.给出一种根据推理过程中规则的出现频率决定其是否保留,从而实现规则数量的动态变化的方法,证明了动态变化过程中规则的数量不会无限增加.实例表明此法是比较有效的.     

Maintaining discovered frequent itemsets: Cases for changeable database and support

Mining frequent itemsets from large databases has played an essential role in many data mining tasks. It is also important to maintain the discovered frequent itemsets for these data mining tasks when the database is updated. All algorithms proposed so far for the maintenance of discovered frequent itemsets are only performed with a fixed minimum support,which is the same as that used to obtain the discovered frequent itemsets. That is, users cannot change the minimum support even if the new results are unsatisfactory to the users. In this paper two new complementary algorithms, FMP (First Maintaining Process) and RMP (Repeated Maintaining Process), are proposed to maintain discovered frequent itemsets in the case that new transaction data are added to a transaction database. Both algorithms allow users to change the minimum support for the maintenance processes. FMP is used for the first maintaining process, and when the result derived from the FMP is unsatisfactory, RMP will be performed repeatedly until satisfactory results are obtained. The proposed algorithms re-use the previous results to cut down the cost of maintenance. Extensive experiments have been conducted to assess the performance of the algorithms. The experimental results show that the proposed algorithms are very resultful compared with the previous mining and maintenance algorithms for maintenance of discovered frequent itemsets.     

一种关联规则挖掘的优化算法

文中基于chi-square检验、有趣度及信息增益理论,给出了一种挖掘优化关联规则的算法。该算法将冗余的规则分为两个部分:一部分规则缺乏统计的相关性,而另外一部分规则不满足“新奇的”要求。实验结果表明算法可以有效地去除冗余规则并提高挖掘效率。     

DualMiner: A Dual-Pruning Algorithm for Itemsets with Constraints

Recently, constraint-based mining of itemsets for questions like find all frequent itemsets whose total price is at least $50 has attracted much attention. Two classes of constraints, monotone and antimonotone, have been very useful in this area. There exist algorithms that efficiently take advantage of either one of these two classes, but no previous algorithms can efficiently handle both types of constraints simultaneously. In this paper, we present DualMiner, the first algorithm that efficiently prunes its search space using both monotone and antimonotone constraints. We complement a theoretical analysis and proof of correctness of DualMiner with an experimental study that shows the efficacy of DualMiner compared to previous work.     

Free-Sets: A Condensed Representation of Boolean Data for the Approximation of Frequency Queries

Given a large collection of transactions containing items, a basic common data mining problem is to extract the so-called frequent itemsets (i.e., sets of items appearing in at least a given number of transactions). In this paper, we propose a structure called free-sets, from which we can approximate any itemset support (i.e., the number of transactions containing the itemset) and we formalize this notion in the framework of -adequate representations (H. Mannila and H. Toivonen, 1996. In Proc. of the Second International Conference on Knowledge Discovery and Data Mining (KDD'96), pp. 189–194). We show that frequent free-sets can be efficiently extracted using pruning strategies developed for frequent itemset discovery, and that they can be used to approximate the support of any frequent itemset. Experiments on real dense data sets show a significant reduction of the size of the output when compared with standard frequent itemset extraction. Furthermore, the experiments show that the extraction of frequent free-sets is still possible when the extraction of frequent itemsets becomes intractable, and that the supports of the frequent free-sets can be used to approximate very closely the supports of the frequent itemsets. Finally, we consider the effect of this approximation on association rules (a popular kind of patterns that can be derived from frequent itemsets) and show that the corresponding errors remain very low in practice.     

A Web-Based Knowledge Network for Supporting Emerging Internet Applications

Although the Internet and the World Wide Web technologies have gained a tremendous amount of popularity among people and organizations, the network that these technologies created is not much more than a multimedia data network. It provides tools and services for people to browse and search for data but does not provide the facilities for automatically delivering the relevant information for supporting decision–making to the right people or applications at the right time. Nor does it provide the means for users to enter and share their knowledge that would be useful for making the right decisions. In this work, we introduce the concept of a Web–based knowledge network, which allows users and organizations to publish, not only their multimedia data, but also their knowledge in terms of events, parameterized event filters, customizable rules and triggers that are associated with their data and application systems. Operations on the data and application systems may post events over the Internet to trigger the processing of rules defined by both information providers and consumers. The knowledge network is constructed by a number of replicable software components, which can be installed at various network sites. They, together with the existing Web servers, form a network of knowledge Web servers.