摘要 :
Negative sequential patterns (NSP) refer to sequences with non-occurring and occurring items, and can play an irreplaceable role in understanding and addressing many business applications. However, some problems occur after mining...
展开
Negative sequential patterns (NSP) refer to sequences with non-occurring and occurring items, and can play an irreplaceable role in understanding and addressing many business applications. However, some problems occur after mining NSP, the most urgent one of which is how to select the actionable positive or negative sequential patterns. This is due to the following factors: 1) positive sequential patterns (PSP) mined before considering NSP may mislead decisions; and 2) it is much more difficult to select actionable patterns after mining NSP, as the number of NSPs is much greater than PSPs. In this paper, an improved method of pruning uninteresting itemsets to fit for a selecting actionable sequential pattern (ASP) is proposed. Then, a novel and efficient method, called SAP, is proposed to select the actionable positive and negative sequential patterns. Experimental results indicate that SAP is very efficient in the selection of ASP. To the best of our knowledge, SAP is the best method for the selection of actionable positive and negative sequential patterns.
收起
摘要 :
Negative sequential patterns (NSPs), which focus on nonoccurring but interesting behaviors (e.g. missing consumption records), provide a special perspective of analyzing sequential patterns. So far, very few methods have been prop...
展开
Negative sequential patterns (NSPs), which focus on nonoccurring but interesting behaviors (e.g. missing consumption records), provide a special perspective of analyzing sequential patterns. So far, very few methods have been proposed to solve for NSP mining problem, and these methods only mine NSP from positive sequential patterns (PSPs). However, as many useful negative association rules are mined from infrequent itemsets, many meaningful NSPs can also be found from infrequent positive sequences (IPSs). The challenge of mining NSP from IPS is how to constrain which IPS could be available used during NSP process because, if without constraints, the number of IPS would be too large to be handled. So in this study, we first propose a strategy to constrain which IPS could be available and utilized for mining NSP. Then we give a storage optimization method to hold this IPS information. Finally, an effcient algorithm called Effcient mining Negative Sequential Pattern from both Frequent and Infrequent positive sequential patterns (e-NSPFI) is proposed for mining NSP. The experimental results show that e-NSPFI can effciently find much more interesting negative patterns than e-NSP.
收起
摘要 :
Mining negative sequential patterns (NSP) has been an important research area in data mining and knowledge discovery and it is much more challenging than mining positive sequential patterns (PSP) due to the computational complexit...
展开
Mining negative sequential patterns (NSP) has been an important research area in data mining and knowledge discovery and it is much more challenging than mining positive sequential patterns (PSP) due to the computational complexity and search space. Only a few methods have been proposed to mine NSP and most of them only use single minimum support, which implicitly assumes that all items in the database are of the same nature or of similar frequencies in the database. This is often not the case in real-life applications. There are several methods to mine sequential patterns with multiple minimum supports (MMS), but these methods only consider PSP and do not handle NSP. So in this paper, we propose a new method, called e-msNSP, to mine NSP with multiple minimum supports. We also solve the problem of how to set up the minimum support to a sequence with negative item(s). E-msNSP consists of three major steps: (i) using the improved MS-GSP method to mine PSP with multiple minimum supports and storing all positive sequential candidates' (PSC) related information simultaneously; (ii) using the same method in e-NSP to generate negative sequential candidates (NSC) based on above mined PSP; (iii) calculating the support of these NSC based only on the corresponding PSP and then getting NSP. To the best of our knowledge, e-msNSP is the first method to mine NSP with MMS and does not impose strict constraints. Experimental results show that the e-msNSP is highly effective and effcient.
收起
摘要 :
Recently, negative sequential patterns (NSP) (like missing medical treatments) mining is important in data mining research since it includes negative correlations between item sets, which are overlooked by positive sequential patt...
展开
Recently, negative sequential patterns (NSP) (like missing medical treatments) mining is important in data mining research since it includes negative correlations between item sets, which are overlooked by positive sequential pattern mining (PSP) (for instance, utilization of medical service). Yet, discovering the NSP is very complex than finding PSP because of the important problem complexity occurred by high computational cost, non-occurring elements, as well as huge search space in evaluating NSC, and most of the NSP based existing works are inefficient. Therefore, this paper intends to propose a fast NSP mining algorithm for the disease prediction model. This model includes Data normalization, Data separation based on labels, and Pattern recognition phases. In the midst of data separation, the maximum occurring data is optimally selected using a new algorithm that hybridizes the FireFly (FF) algorithm and Grey Wolf Optimization (GWO). This proposed Firefly induced Grey Wolf optimization (F-GWO) algorithm automatically selects the maximum occurring information as per the PSP support. The proposed model is compared over other conventional methods with varied measures. Especially, the computation cost of our model is 46.87%, 6.27%, 9.37%, 2.76%, and 66.62% better than the existing GA, ABC, PSO, FF, and GWO models respectively.
收起
摘要 :
In data mining field, sequential pattern mining can be applied in divers applications such as basket analysis, web access patterns analysis, and quality control in manufactory engineering, etc. Many methods have been proposed for ...
展开
In data mining field, sequential pattern mining can be applied in divers applications such as basket analysis, web access patterns analysis, and quality control in manufactory engineering, etc. Many methods have been proposed for mining sequential patterns. However, conventional methods only consider the occurrences of itemsets in customer sequences. The sequential patterns discovered by these methods are called as positive sequential patterns, i.e., such sequential patterns only represent the occurrences of itemsets. In practice, the absence of a frequent itemset in a sequence may imply significant information. We call a sequential pattern as negative sequential pattern, which also represents the absence of itemsets in a sequence. The two major difficulties in mining sequential patterns, especially negative ones, are that there may be huge number of candidates generated, and most of them are meaningless. In this paper, we proposed a method for mining strong positive and negative sequential patterns, called PNSPM. In our method, the absences of itemsets are also considered. Besides, only sequences with high degree of interestingness will be selected as strong sequential patterns. An example was taken to illustrate the process of PNSPM. The result showed that PNSPM could prune a lot of redundant candidates, and could extract meaningful sequential patterns from a large number of frequent sequences.
收起
摘要 :
A growing movement in the United States and around the world involves promoting the advantages of conducting an eyewitness lineup in a sequential manner. We conducted a large study (N = 2,529) that included 24 comparisons of seque...
展开
A growing movement in the United States and around the world involves promoting the advantages of conducting an eyewitness lineup in a sequential manner. We conducted a large study (N = 2,529) that included 24 comparisons of sequential versus simultaneous lineups. A liberal statistical criterion revealed only 2 significant sequential lineup advantages and 3 significant simultaneous advantages. Both sequential advantages occurred when the good photograph of the guilty suspect or either innocent suspect was in the fifth position in the sequential lineup; all 3 simultaneous advantages occurred when the poorer quality photograph of the guilty suspect or either innocent suspect was in the second position. Adjusting the statistical criterion to control for the multiple tests (.05/24) revealed no significant sequential advantages. Moreover, despite finding more conservative overall choosing for the sequential lineup, no support was found for the proposal that a sequential advantage was due to that conservative criterion shift. Unless lineups with particular characteristics predominate in the real world, there appears to be no strong preference for conducting lineups in either a sequential or a simultaneous manner.
收起
摘要 :
Threshold group testing introduced by Damaschke (2006) is a generalization of classical group testing where a group test yields a positive (negative) outcome if it contains at least u (at most l) positive items, and an arbitrary o...
展开
Threshold group testing introduced by Damaschke (2006) is a generalization of classical group testing where a group test yields a positive (negative) outcome if it contains at least u (at most l) positive items, and an arbitrary outcome for otherwise. Motivated by applications to DNA sequencing, group testing with consecutive positives has been proposed by Balding and Torney (1997) and Colbourn (1999) where n items are linearly ordered and up to d positive items are consecutive in the order. In this paper, we introduce thresholdconstrained group tests to group testing with consecutive positives. We prove that all positive items can be identified in 「log2(「n/u」?1)」+2「log2(u+2)」+「log2(d?u+1)」?2 tests for the gap-free case (u = l+1) while the information-theoretic lower bound is 「log2 n(d? u+1)」?1 when n ≥ d+u?2 and for u = 1 the best adaptive algorithm provided by Juan and Chang (2008) takes at most 「log2 n」+「log2 d」tests.Wefurther show that the case with a gap (u > l + 1) can be dealt with by the subroutines used to conquer the gap-free case.
收起
摘要 :
Sequential products of quantum measurements are defined and studied. Two types of measurement equivalence are considered and their relationships with compatibility and the sequential product are discussed. It is shown that a measu...
展开
Sequential products of quantum measurements are defined and studied. Two types of measurement equivalence are considered and their relationships with compatibility and the sequential product are discussed. It is shown that a measurement A is sharp if and only if A is equivalent to the sequential product of A with itself. Refinements of measurements are defined and it is shown that they produce a partial order on the set of measurements. Lattice properties of this partially ordered set are briefly discussed. Finally we consider convex combinations and conditioning for quantum measurements.
收起
摘要 :
We investigated the contribution of cognitive ability and affect to age differences in sequential decision making by asking younger and older adults to shop for items in a computerized sequential decision-making task. Older adults...
展开
We investigated the contribution of cognitive ability and affect to age differences in sequential decision making by asking younger and older adults to shop for items in a computerized sequential decision-making task. Older adults performed poorly compared to younger adults partly due to searching too few options. An analysis of the decision process with a formal model suggested that older adults set lower thresholds for accepting an option than younger participants. Further analyses suggested that positive affect, but not fluid abilities, was related to search in the sequential decision task. A second study that manipulated affect in younger adults supported the causal role of affect: Increased positive affect lowered the initial threshold for accepting an attractive option. In sum, our results suggest that positive affect is a key factor determining search in sequential decision making. Consequently, increased positive affect in older age may contribute to poorer sequential decisions by leading to insufficient search.
收起
摘要 :
Liking and wanting are two foundational processes underlying the individual's reward system. Whereas the differences between liking and wanting have been studied extensively in the context of substance addiction, there have been f...
展开
Liking and wanting are two foundational processes underlying the individual's reward system. Whereas the differences between liking and wanting have been studied extensively in the context of substance addiction, there have been few empirical studies of their manifestation in ordinary, nonaddictive contexts of behavior. In particular, previous research showed a temporal divergence of liking and wanting over repeated exposures to drugs; however, the temporal progression of liking versus wanting in response to ordinary stimuli remains less understood. This research tests the temporal divergence of liking versus wanting responses in a prevalent domain-namely, in response to persuasive messages-using a highly powered field experiment involving over 1,000 real-life stimuli and 100,000 subjects. Subjects were exposed to sequences of TV advertisements in random orderings and indicated how much they liked the persuasive messages and wanted the promoted items. We found, not surprisingly, an overall positive correlation whereby greater message liking was associated with greater wanting of the item. However, underlying the overall correlation, we found a sharp divergence with respect to the serial progression of liking versus wanting. Specifically, message liking was highest early in the sequence, whereas wanting of the promoted item was highest late in the sequence. We posit and provide evidence that this divergence is due to an appetizing effect for wanting, in contrast to a habituation and reference adaptation effect for liking. We discuss the implications of our results for theories of liking versus wanting for addictive as well as ordinary substances.
收起