| e707004304 |
 Wysłany: Sob 10:41, 15 Sty 2011 Temat postu: Rough Set Based Information System for Incomplete |
|
Incomplete based on rough sets A Complete Information System
To measure the similarity between two objects, we need to be quantified. Definition lO: Suppose A = port l = 1,2, ..., P is the set of attributes, each attribute value uniform distribution, for any one if_A, its range is =,, ...,}, I () that the object the property port values, the object in the property the mouth and f similar to the probability P (i √) as: P (i √) = (port () ≠ ()) eight (port () ≠ Pán c) Eight (mouth () ≠ Pán c) (I () = Pán c) eight (I () = Pán c) (I () = Pán c) eight (() ≠ Pán c) V {(() ≠ Pán c) VIII (I () = Pán c) (I () = 'I (,)) eight (() ≠) eight (port () ≠) to the object and the attribute set A, the similarity: P (,: fP (√), ≠., 【l,ghd piastre,. 『Based on the above definition, we come to the right ROUSTIDA algorithm is improved, the improved algorithm for missing property values of objects exist, using the maximum degree of similarity with the object that has the highest property values missing to fill its capacity to fill an object, reducing the number of over-use of information systems to accelerate the calculation speed. as follows: Input: incomplete information system S. = (, A,), Output: a complete information system S ' = (, A,). Step 1: Calculate the discernibility matrix and the initial expansion gAS?, MOS.; that t = 0. Step 2: (1) For all if_MOS ', calculated as:. (2) production Js \For the MOS ', shell 0 (t) = population ( ,= 1,2, ..., P. For if_MOS', for all if_gAS: for recycling: If lLS: = 1l, set fif_LS:,asics australia, if the mouth (ft ) =: Ic, then the mouth ('t ¨) =; or oral (t \≠) Eight (port (f2) ≠), then (t ¨) =: Ic; Otherwise, if there is fif_LS:, satisfy () ≠: Ic, then the mouth (t) = (,ugg italia, t); otherwise, (t ¨ ) =: Ic. (3) If S \, property values are frequently used to fill missing values and other methods to deal with. Step 4: End. 3 Table 1 for examples of intelligence information systems is not complete S.,christian louboutin pas cher, one mouth, mouth, mouth, mouth to condition attributes, mouth if_ 0,1,2,3, mouth if_ 0,1,2,3, rough set-based information system is not perfect completeness of the information port E 0,1,2, port E 0 , 1. mouth as decision attribute, Oral E 0,1. that missing data. ROUSTIDA algorithm used to produce information systems to fill the Js., Table 2, that is,tory burch outlet, calculating this table, the algorithm satisfies the termination condition , can not be filled. The use of improved algorithm, without step 3 with the other methods can be incomplete intelligence information systems Js. completions, completion of a Information System Js: Table 3. Table 1 does not completely Intelligence Information Systems | s. Table 2 Information Systems | s table 3 full of intelligence and information systems | s: the Js: with the Js. comparison, it is worth noting that some of the objects with missing data (such as), although two algorithms get the same results to fill, but fill the result of the use ROUSTIDA algorithms similar to object because of missing values on the properties of the resulting conflict-free (.,, in port: The port property values 2 and 1, respectively), while Improved results are obtained by the algorithm to fill the largest one-time use of the similarity of objects (..) Table 4 is filled in the use of improved algorithms for the first time during the calculation of the similarity of the object table, as one by one to fill the missing values, only This table will cause the change of local elements. Table 4, the similarity of the object table |
|
