Normalized-Mutual-Information-Based Mining Method for Cascading Patterns

Abstract

Abstract:A cascading pattern is a sequential pattern characterized by an item following another item in order. Recent research has investigated a challenge of dealing with cascading patterns, namely, the exponential time dependence of database scanning with respect to the number of items involved. We propose a normalized-mutual-information-based mining method for cascading patterns (M 3 Cap) to address this challenge. M 3 Cap embeds mutual information to reduce database-scanning time. First, M 3 Cap calculates the asymmetrical mutual information between items with one database scan and extracts pair-wise related items according to a user-specified information threshold. Second, a one-level cascading pattern is generated by scanning the database once for each pair-wise related item at the quantitative level. Third, a recursive linking-pruning-generating loop generates an (m + 1)-level-candidate cascading pattern from m-dimensional patterns on the basis of antimonotonicity and non-additivity, repeating this step until no further candidate cascading patterns are generated. Fourth, meaningful cascading patterns are generated according to user-specified minimum evaluation indicators. Finally, experiments with remote sensing image datasets covering the Pacific Ocean demonstrate that the computation time of recursive linking and pruning is significantly less than that of database scanning; thus, M 3 Cap improves performance by reducing database scanning while increasing intensive computing.

An English-Chinese Machine Translation and Evaluation Method for Geographical Names

“…To solve the above problems, this paper establishes an English-Chinese machine translation and evaluation model for geographical names based on the word-formation characteristics and attribute information of English geographical names, the theoretical knowledge related to machine learning [13] and the geographical name translation standards of China [14]. The basic thinking is as follows: first, all toponymic data are divided into groups according to the category attributes; then, based on pointwise mutual information [15], common phrases of different categories of the corpus are calculated and explored [16], and the data structure of a directed acyclic map [17] is used to extract the geographical name template. In the process of translation, the same category of the template is used to nest matching geographical names and split their structures completely to generate a lexical structure tree [18,19].…”

Spatiotemporal event sequence discovery without thresholds

“…In the spatiotemporal frequent pattern mining literature, the term sequence (or its derivatives such as sequence patterns, sequential patterns) is used for identifying different types of knowledge from spatiotemporal data. These include sequences of locations frequently visited by spatiotemporal objects [ 14 , 20 ], partially- or totally-ordered sequences of event types whose instances follow each other [ 5 , 13 , 24 , 34 , 35 , 49 ] (these are also referred to as couplings in [ 42 ]), sequences of semantic annotations from semantic trajectories [ 51 ], temporal sequences of ordered spatial itemsets (called spatio-sequences) [ 40 ], and sequences of spatiotemporal association rules [ 47 ].…”