Efficient MaxCount and threshold operators of moving objects

Abstract

Calculating operators of continuously moving objects presents some unique challenges, especially when the operators involve aggregation or the concept of congestion, which happens when the number of moving objects in a changing or dynamic query space exceeds some threshold value. This paper presents the following six d-dimensional moving object operators: (1) MaxCount (or MinCount), which finds the Maximum (or Minimum) number of moving objects simultaneously present in the dynamic query space at any time during the query time interval. (2) CountRange, which finds a count of point objects whose trajectories intersect the dynamic query space during the query time interval. (3) ThresholdRange, which finds the set of time intervals during which the dynamic query space is congested. (4) ThresholdSum, which finds the total length of all the time intervals during which the dynamic query space is congested. (5) ThresholdCount, which finds the number of disjoint time intervals during which the dynamic query space is congested. And (6) ThresholdAverage, which finds the average length of time of all the time intervals when the dynamic query space is congested. For these operators separate algorithms are given to find only estimate or only precise values. Experimental results from more than 7,500 queries indicate that the estimation algorithms produce fast, efficient results with error under 5%.

A Dynamic Data Structure to Efficiently Find the Points below a Line and Estimate Their Number

“…Selectivity estimation is an important problem in spatial and spatio-temporal database querying because the estimate guides the evaluation of the query [4,6,7]. Selectivity estimation for moving point objects was considered by Anderson and Revesz [8], Choi and Chung [9] and Sun et al [10], who described different algorithms for estimating the number of moving points that will be within a specific box area or hyperbox region at a future time t.…”

Classification integration and reclassification using constraint databases

“…= 0 This query returns ''Disease'' if the patient is sick. Constraint databases, which were initiated by the original article of Kanellakis et al [16], have many applications ranging from spatial databases [17,18] through moving objects [19,20] to epidemiology [21]. The original constraint database model was extended recently to labelled object-relational constraint databases (LORCDB) by Gó mez-Ló pez et al [22] and applied to model-based diagnosis.…”

“…(16) platelets per cubic ml/1000, (17) prothrombin time in seconds, (18) status (0 = alive, 1 = transplant, or 2 = dead), (19) drug (1 = D-penicillamine or 2 = placebo), and (20) histologic stage of the disease (1, 2, 3, 4).…”

“…We generated the following two subsets from the original data set: DISEASE with features (3,4,5,7,8,9,10,13,14,16,17,20) DRUG with features (3,4,6,7,8,9,10,11,13,16,17,19) Results: In both subsets, we used the first eleven features to predict the last feature. Fig.…”

Temporal Data Classification Using Linear Classifiers

“…Constraint databases, which were initiated by Kanellakis et al [12], have many applications ranging from spatial databases [21,6] through moving objects [10,2] to epidemiology [20]. However, only Geist [8] and Johnson et al [11] applied them to classsification problems.…”