4.2. Illustration
Here are the results of running each step on the sample database MEFIM algorithm in Table 2 with minutil = 35
Step 1. Initialize X = ∅
Step 2. lu (X, i) with i∈I and X = ∅
Step 3. From lu (X, i) we have Secondary (X) = {a, b, c, d, e}
Step 4. Arrange Secondary (X) increases with lu (X, i) Secondary (X ) = {e, c, d, b, a}
Step 5. Scan the data and delete elements not belonging to Secondary (X)
Step 6. Arrange the transaction of D (see results in Figure 3)
step 7. rubber (X, i) with i∈Secondary (X) and X = ∅ (see Figure 2)
step 8. At Primary (X) = {e, c, d}
recursive function step 9. Call Search (X = ∅, D, Primary (X) = {e, c, d}, Secondary (X) = {e, c, d, b, a}, minutil = 35)
- β = X ∪ {e} = {f}
- turn transactions browse T1, T5, T3, T4, T2 to calculate u (β) = 8.8 + 4 = 12.8, and create projections on D_β database. Results shown in Figure 4
- As lu (β, i) and rubber (β, i) with i∈Secondary (X)
- At the threshold minutil = 35, inferred Primary (β) = {d} and Secondary ( β) = {d, b, a}
+ same recursion Search function (X = {e}, D_X, Primary (X) = {d}, Secondary (X) = {d, b, a}, minutil = 35) we have u (β) = 23.1 β = {ed} the ones shown in Figure 5
+ do the same recursive function Search to continue to expand the set {ed} have been highly useful set {1 edba u} ({edba}) = 38.9.
- Similar to browse turn 5 transactions for each case extended with = {c} β {c} = u 8.1, β = {d} with u ({d }) = 20.8, and we do not find more helpful is set high in 2 cases. See Figure 6 and 7
đang được dịch, vui lòng đợi..
