Communications of the ACM, vol 40, no 3, pp 56-58, 1997. [1 P. Resnick and H. R. Varian, " Recommender [2]L. Terveen and w. Hill, "Beyond Recommender Systems: Helping cople Help Each Othe HCI in the New Millennium, pp. 487- 3]J. Schafer, J Konstan, and J. Riedl, "Recommender Systems in E- the Ist ACM conference on Electronic commerce, pp 158-166, NY: USA, 1999 14 M. Balabanovic and Y. Shoham, "FAB: Content-based, Collaborative Recommendation, Communications of the ACM, vol 5]R. Burke, "Hybrid Web Recommender Systems, "Lecture Notes in The number of items rated over 80 Heidelberg, 200 Figure 5. The experiment result of 10 users for 7 pages for Collaborative Demographic Filtering, " Artificial Intelligence Review, voL 13,no5- g 6,pp.394408,1999 recommender system. We did not consider results of the first 7 M. Mitchell, dn Introduction to Genetic Algorithm, MIT Press, 1998 items on this page than other pages. Besides, the users need [8] D. E Gold berg and J.H. Holland, "Genetic Algorithms and Ma page since users tend to rate too much higher values for to become familiar with the rating mechanism of our system. [9 H. Takagi, "Interactive Evolutionary Computation: Fusion of the Also, we measured the number of items that were rated over 80 for each pages because these items have more impact to Proceedings of the IEEE, vol. 89, pp. 1275-1296, 200 trace users'preference than others. As seen in the figure, the [10] D. Thierens and D. Goldberg, "Elitist Recombination: An Integrated (i.e, as the generation goes by/ l e page number scores gradually improve as the Conference on Evolutionary Computation, pp. 508-512, 1994 re satisfied with the recommendation made by our [J J. E Crow and m ra,"Efficiency of Truncation system. In other words, the results prove the efficacy of our Proceedings of the National Academy of Sciences of the f America, voL. 76, no. 1, pp 396- recommender system in detecting and tracing the preference [12J F. Herrera, M. Lozano, and J. L. Verdegay, " Tackling Real-coded of each user ms:Operators and Tools for Behavioural Analysis, Artificial Intelligence Review, voL. 12, pp. 265-319, 1998 V CONCLUSION In this paper, we presented a novel recommender system for or music data. Our proposed system is able to recognize the preference of users and then adaptively recommend music tracks appropriate for their present atmosphere To construct the system, we incorporated the main interactive genetic algorithm-based engine with the content- based filtering method. First, the unique features of each music track are extracted. We then apply the interactive genetic algorithm to obtain the most appropriate tracks to recommend to users. The proposed system enables users to evaluate the fitness of each recommended item according to eir preferences. Therefore, our system can gradually adapt its recommendation with the subjective decision of each user. The experimental results exhibited that the average scores, which are objectively collected by means of user aluations, increases by degrees as th denotes that the proposed system can help the users obtain various music tracks suitable with their own preference. Therefore, we believe that the recommender system retains enough potential to be implemented and applied to other platforms(e.g, standalone program, web service and mobile device) This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government(MEST)(No. 2009-0066229 Volume 6] 2010 2nd International Conference on Computer Engineering and Technology V6-417:F3 85  >/25 :3 F .  5     55    3%   5           -    5 53  <  .     3 .   < .- H/ 5< -  5 ;5  3   -    5-   5  < 5*33  <+3(   .  .     <   3( .5-        5  3 D3 $76$#, (76 ( 55.5  -    3755 < = 5       5-     &5555  53       .  5    -  <  .   <   3 :   9     &  83 %  55   -     <   555 &     3  55   <  -      5 3   5    .<I- 3  85    8<   -  .  <I-  <     -  <  .3(   55   5  < -  & < .  . 5 3  . <-         5   <  5     55   5 *33  5 .<-  < -+3 $67%#!!6  .& . 55 <  6   :    *6:+     <   -  *! +*63?//B//11??B+3 !:!!6$! @>A L3  &   3 3 D  P    Q   #  -30/ 3C553F1RFH>BB23 @?A #3 -  %3  P      E 5  L55!7Q . "# / 0  553 0H2R F/B %?//>3 @CA "3  "3      "3  P      !  Q &   # (            553>FHR>116JE, >BBB3 @0A 3 < -;   J3   P:E $  < $<-    Q   #    -3 0/ 3C55311R2?>BB23 @FA 3 & P<%<     Q 1   /    %     -3 0C?> 553 C22R0/H 5   E <?//23 @1A 3 "3 L==  P : .& $<- $  <    5: Q"     0-3>C 3F 1553CB0R0/H>BBB3 @2A 3"   !   #( L>BBH3 @HA 3!3< "33 P     #  Q# 1 -3C 3?C553BFRBB>BHH3 @BA 3 & P( - !-  $ 5 E :   $5<  !$ 75 =     !- Q &   # "-3HB553>?2FR>?B1?//> @>/A 3   3<P! <  E (    <  Q  &    #   "        % 553F/HRF>?>BB03 @>>A "3:3$. 3 P!     Q &   # /      # 2    -321 3>553CB1RCBB>B2B3 @>?A :3  3 #=    "3 #3 D P &     E75  - Q "     0-3>?553?1FRC>B>BBH3 [Volume 6] 2010 2nd International Conference on Computer Engineering and Technology V6-417