Input: $R\left(A_i,A_2,\cdots ,A_m\right)$ : The relation; $Q=\left\{Q_1,Q_2,\cdots ,Q_n\right\}$ : The set of user queries that will run on R; cut-off: Cut-off Average Similarity .

Output: C: The set of materialized views.

Step 1: Construct the m × n Attribute Usage Matrix where m is the number of attributes and n is the number of queries.

Step 2: Construct the n × n Attribute Similarity Matrix from the Attribute Usage Matrix.

/* In step 3 the Graphs & the corresponding clusters are constructed */

Step 3: ∀ (A_i, A_j) ∈ Attribute Similarity Matrix, identify the attribute pair similarity values so that J (A_i, A_j) ≥ cut-off.

Store these values in a list L in descending order.

Store the number of elements in L into a variable count.

a) If count = 0 then go to step 5. /* The algorithm fails to draw any graph. Therefore, no materialized view can be constructed */

b) If count > 0 then Repeat until list L is empty

i) Pop the ith (1 ≤ I ≤ count) element from L. Store the value of the element into avariable x (0 ≤ x ≤ 1).

ii) ∀ Attribute pairs (A_i, A_j) ∈ Attribute Similarity Matrix

If J (A_i, A_j) = x, then G_i ← G_i ∪ A_i ∪ A_j /* Pair of attributes is inserted

in G_i, where G_i represents the i^th graph for i^th highest similarity value. */

iii) C_i← G_i /*C_i represents the i^th cluster for i^th highest similarity value. */

iv) C ← C ∪ C_i /* Each cluster C_i is added to the set of clusters C */

/* In step 4 the clusters are validated and materialized views are formed*/

Step 4: ∀ C_i ∈ C, Calculate (C_i)_avg by equation:

${\left(C_i\right)}_{avg}=\frac{{\displaystyle {\sum }_{i=1}^{\left(\begin{array}{c}n\\ 2\end{array}\right)}J\left(A_i;A_j\right)}}{\left(\begin{array}{c}n\\ 2\end{array}\right)}$ where i # j

If ((C_i)_avg < cut-off then C ← C − C_i

/* C represents the set of materialized views where each cluster C_i represents a valid materialized view */

Step 5: End.