Thus, it can be seen in Table 3 and Table database 4 is similar in structure, should be able to modify the HUI mining algorithms available in this case.
The convention 1. The data of all illustrations press onward, when it comes to trading is said to have converted the transaction form as in Table 3.
Definitions 4. the value of the set X useful in transaction T, denoted u (X, T), useful is the total value of the elements of X are included in the transaction T and u (x, T) = Σ_ (i∈X∧X⊆T) ▒ 〖u (i, T)〗 [6]
the definition 5. the value of the set X utility, denoted u (X), is the total value of X useful in all transaction T contains X on DB or u (X) = Σ_ (T∈DB∧X ⊆T) ▒ 〖u (T).〗 [6]
definition 6. Given a minimum threshold minutil useful, set X is called useful set high if the value of X is not useful or is less than the threshold u ( X) ≥minutil. [6]
For example: u ({ab}, T2) = u (a, T2) + u (b, T2) = 4 + 1.9 = 5.9, and u ({ab}) = u ({ab}, T2 ) + u ({ab}, T4) + u ({ab}, T5) = 5.9 + 9.9 + 13.2 = 34.
If minutil {ab} = 30, it is highly useful practice, as opposed to minutil = 40, then { ab} is not highly useful set.
Definitions 7. the value of the transaction T helpful, religious symbols (T), is the total value of the parts can be useful in T or tu (T) = Σ_ (i∈T) ▒ 〖u (i, T)〗 and usefulness of the DB value is the total value of transactions helpful in DB [6].
for example: tu (T3) = u ({a}, T3) + u ({c}, T3) + u ({d}, T3) = 4.4 + 2.2 + 5.5 = 12.1
definition 8. Weighted useful transactions set X, denoted TWU (X), is useful total value of all transactions containing X on DB or TWU (X) = Σ_ (T∈DB∧X⊆T) ▒ 〖tu (T)〗 [6].
for example, TWU ({ e}) = tu (T2) + tu (T4) = 17.4 + 24.2 = 41.6
Definitions 9. Call ≻ is allowed to order the elements of the first set according to TWU. Remaining useful value of X in the transaction T, ru symbols (X, T) is the total value of the following elements useful X in T, or ru (x, T) = Σ_ (i∈T ∧ ∀ x ∈ x ≻x i) ▒ 〖u (i, T)〗. [3]
For example: ru ({a}, T3) = u ({c}, T3) + u ({d}, T3) = 2.2 + 5.5 = 7.7
Definitions 10. For the first set of elements rated order by ≻, and set X, the set of elements of X expansion is defined as follows E (X) = {z ┤ | z∈I∧z≻x∀x∈X} [11]
the definition 11. for transactions T and set X, the projection of X on the transactions set T is defined as T_X = {i ┤ | i∈T∧i∈E (X)} [11]
for example: Let X = {b} , review the order allowing the ≻e ≻d a≻b ≻c〗 〖_X = ∅ T1, T2〗 〖_X = {a}
definition 12. Given a database D and set X, the projection of the set X on D is defined as follows D_X = {T_X ┤ | T∈D∧T_X ≠ ∅} [11]
For example: Let X = {c}, the order allows a≻b ≻c review ≻d ≻e, D_X = {〖〗 _X T1, T2 〖〗 _ (X) , 〖〗 _X T3, T5〗 〖_X} = {{b}, {ab}, {a}, {ab}}
definition 13. Let set X, z∈E element (X) and the useful value local (x, z) is calculated as follows lu (x, z) = Σ_ (T⊃ (X ∪ {z})) ▒ 〖[u (x, T) +〗 ru (X, T)] [11]
For example. Let X = {a}, lu (X, c) = (u (X, T2) + ru (X, T2)) + (u (X, T3) + ru (X, T3)) + (u (X T5) + ru (X, T5)) = 17.4 + 12.1 + 21.6 = 51.1
Characteristics 1. set X, z∈E (X), if lu (X, z)Definition 14. Let X and element sets z∈E (X)), utility value on the tributaries and set X is su z (x, z) = Σ_ (T⊃ (X ∪ {z})) ) ▒ 〖[u (α, T) + u (z, T) +〗 Σ_ (i∈T⋀i∈E (α⋃ {z})) ▒ 〖u (i, T)]〗 [11]
Example. Let X = {a}, rubber (X, c) = (u ({a}, T2) + u ({c}, T2) + u ({d}, T2) + u ({e}, T2) ) + (u ({a}, T3) + u ({c}, T3) + u ({d}, T3)) + (u ({a}, T5) + u ({c}, T5) + u ({f}, T5)) = 15.5 + 12.1 + 12 = 39.6.
Characteristics 2. X and z∈E set (X), if su (x, z)15. Give the definition set X, major elements and sub-elements (Primary, Secondary item) is defined as follows Primary (X) = {z ┤ | z∈E (X) ∧su (x, z) ≥minutil } and Secondary (X) = {z ┤ | z∈E (X) ∧lu (x, z) ≥minutil}. [11]
For example. Continuing the example in the definition of 13 and 14, given minutil = 40, then X = {a} is the first sub-element but not the main element, but with minutil = 30, then X = {a} is both the element both key elements.
definition 16. Given two transactions Ta, Tb contains elements corresponding {i1, i2, ..., im} and {j1, j2, ..., jn}. Ta and Tb are called homogeneous or Ta = Tb, provided the following conditions n = m and ∀ k∈ [1, n], ik = jk [11].
For example: Consider the example in the definition to 12, then T2〗 〖_ (X) and T5〗 〖_X be considered uniform for the same results as the {a, b}.
definition 17. for transactions TR1 = TR2 = uniform .... = TRM on D, the transaction is remixed by ∀ i ∈ Tm which T_m, u (i, T_ (m)) = Σ_ (k = 1 ... m) ▒ 〖u (i, T_k)〗.
example: Suppose 〖〗 _ T2 (X) and T5 〖defined〗 _X at 12 is two independent transactions, the two transactions are replaced by 〖T2 '〗 _ (X) with useful internal values the u ({a}, 〖T2 ^ '〗 _ (X)) = u ({a}, T2 〖〗 _ (X)) + u ({a}, 〖〗 T5 _ (X)) = 4 + 3.6 = 7.6 and u ({b}, 〖T2 ^ '〗 _ (X)) = u ({b}, T2 〖〗 _ (X)) + u ({b}, 〖〗 T5 _ (X)) = 1.9 + 9.6 = 11.5
Definitions 18 (the projection combines homogeneous mixing transactions) when the projector set up D X, homogeneous transactions is mixed with a new transaction, symbol CDX. [11].
For example, the projection combine to allow mixed concrete is shown in Figure 5 illustrates an example of MEFIM algorithm.
đang được dịch, vui lòng đợi..
