Figure 2 - uploaded by Ronnie Alves

Content may be subject to copyright.

Source publication

In many applications, data contains structured information that is multidimensional and multilevel
in nature, such as the ones in areas like e-commerce, telecommunications, retail, stocks, or
bioinformatics. Since the last decade, we have been facing several research efforts on Data
Warehousing and OLAP to get better view of multidimensional data,...

## Contexts in source publication

**Context 1**

... we may think as strong candidate dimensions, the date, the product and the store. Further refinements in this model we head us to a retail sales schema as presented in Figure 2. Let us say that the retail sales schema is done and now we are able to answer a couple of questions. ...

**Context 2**

... figure 12 we exemplify such evaluation by allowing Top-K queries over the complete set of alarms. We also verified that the most 10 imperative anomalous situations were raging from 2.76 up to 3.33 concerning its distance function. ...

**Context 3**

... complete view of this labeling task for the entire week is illustrated on Figure 20 (b). As we would expect, the number of new vertices increase regularly when compared with the number of edges (see Figure 20 (a)). ...

**Context 4**

... complete view of this labeling task for the entire week is illustrated on Figure 20 (b). As we would expect, the number of new vertices increase regularly when compared with the number of edges (see Figure 20 (a)). Although, a large number of potential fraud situations were triggered, preliminary feedback given by the fraud analysts has already proved that our methods are a valuable tool to assist them in fraud detection. ...

**Context 5**

... from the classical association rules algorithms, that usually take a flat database to extract From the above definitions we can define our problem of mining evolving data cubes as: Given a base Relation R, which evolves through times instants t to t n+1 , extract interesting inter-dimensional and multilevel patterns DM, on incremental cubing basis. Figure 21. The complete set of partitions provided by the cube structure for Example 1 ...

**Context 6**

... final array is formed by a group of eight partitions. In Figure 21, we show just the first three tuples processed from the inventory fact table of the FoodMart Warehouse provided by Microsoft SQL Server. ...

**Context 7**

... we are working on extending the method to work with complex aggregating functions, in those experiments, just distributive ones were taken into account. Figure 22. Shows the updating effects (dg%, x-axis) against cubing speedup(ic/rc, y-axis). ...

**Context 8**

... Figure 22(a), 22(b) and 22(c) allows observe that in general ic performs much better than rc, specially when the degree of updating is low. We also notice that the speedup increases with respect to the size of the DM. ...

**Context 9**

... have elaborated several experiments (observe the running time on ms) based on the value of the minimum support (%) with respect to different settings of predicates (see Figure 24). We also have constrained the maximum number of levels to explore as 3, and the maximum number of dimensions to 4. Figure 24(a) and 24(c) show the performance figures of mining interesting relations through different thresholds using R-F dataset. ...

**Context 10**

... have elaborated several experiments (observe the running time on ms) based on the value of the minimum support (%) with respect to different settings of predicates (see Figure 24). We also have constrained the maximum number of levels to explore as 3, and the maximum number of dimensions to 4. Figure 24(a) and 24(c) show the performance figures of mining interesting relations through different thresholds using R-F dataset. We can see that increasing the number of hierarchies (levels) doesn't imply high costs on our incremental OLAP Mining. ...

**Context 11**

... other way around, increasing the number of dimensions plays few overheads. Although, we can also save cube computation when exploring the downward property of 3CV measure ( Figure 23). ...

**Context 12**

... figures with R-W dataset are given in Figure 24(b) and Figure 23(b). Again, we can see the effects of mining correlated cuboids when evaluating interesting relations on each dataset. ...

**Context 13**

... figures with R-W dataset are given in Figure 24(b) and Figure 23(b). Again, we can see the effects of mining correlated cuboids when evaluating interesting relations on each dataset. ...

**Context 14**

... 3CV values will provide interesting relations regardless the high overhead provided by the conf measure. We can evidence this tradeoff while OLAP Mining datasets R-F and R-W on Figure 23(a) and Figure 23(b), respectively. Finally, we are able to evaluate our method with respect to update effects on a fact table. ...

**Context 15**

... 3CV values will provide interesting relations regardless the high overhead provided by the conf measure. We can evidence this tradeoff while OLAP Mining datasets R-F and R-W on Figure 23(a) and Figure 23(b), respectively. Finally, we are able to evaluate our method with respect to update effects on a fact table. ...

**Context 16**

... index strategy adopted by our method is a join-index on the combined (foreign key) FK columns in the fact table. We provide an index for each partition (see Figure 21). In fact each cuboid has an index associated to. ...

**Context 17**

... the information is inserted. Figure 25, shows the effects of processing (DAQ) speed up while indexing datasets through different updating scenarios. The little overhead in Figure 25(b) is given by its multilevel properties (ranging from 3 to 4). ...

**Context 18**

... 25, shows the effects of processing (DAQ) speed up while indexing datasets through different updating scenarios. The little overhead in Figure 25(b) is given by its multilevel properties (ranging from 3 to 4). ...

**Context 19**

... partitions -The partition induced by cover equivalence is convex. Cover partitions can be grouped into a M3C class (ji..jn) (Figure 26). Each class in a cover partition has a unique maximal upper bound, and a unique lower bound (Table 17). ...

**Context 20**

... local_3CV value of a class is the maximum local value given by the lower bound cell of each class (Table 17). Figure 26. The M3C lattice formed by the cover partitions of the Example 1. ...

**Context 21**

... the experiments were performed on a 3GHz Pentium IV with 1Gb of RAM memory, running Windows XP Professional. TopKgr-Cube was coded with Java 1.5, and it was performed over synthetic datasets (Table 21) (Figure 32) we want to see the effects of using variance as a criteria for mining Top-K cells: two, (Figure 33) we want to see the effects of iceberg condition while selecting interesting regions; and three, ( Figure 34) the pruning effects by using the heuristics used by our ...

## Similar publications

Binary or multinomial data often occur in agricultural and biological research. Advancements in measurement and video technologies now allow such data to be sequentially recorded through time or space. These data sets, however, can exhibit a serial correlation structure, which in turn, can bias and influence point estimates as well as inferences ma...