Khai phá luật kết hợp gia tăng trên dữ liệu giáo dục

13 HCM Tp HCM ngày 22 tháng 2013 GV Thành i - - : 60.48.01 I II Tìm p III : 02/07/2012 IV V : 21/06/2013 : TS Tp HCM, ngày 21 ii iii T Trong , gia L iv ABSTRACT In recent years, data mining has been applied to education, which leads to a new research area called Educational Data Mining (EDM) In this area, association rule mining is one of the most used techniques Therefore, we have found out research related to the incremental quantitative association rule mining problem on EDM This problem is important to support knowledge discovery from educational data incrementally coming every semester and every year However, it has not yet been investigated much in the existing works So this thesis proposes a quantitative association rule mining approach which is appropriate for an academic credit system which has been popularly applied in the world as well as the incremental problem The mined quantitative rules will give more information to users than the traditional boolean ones As a result, such rules will contribute to the support of educational decision making more conveniently v y Tp HCM, ngày 20 tháng vi M CL C - ii iii iv ABSTRACT - v vi vii ix x - - - - - 10 - 10 12 - 12 -tree - 14 19 vii - 19 27 31 3.1 - 31 33 - 34 40 40 - 40 - 42 42 - - 45 -Growth 46 - 47 - 50 5.1 - 50 - 50 - 51 - 59 62 67 - 67 - 67 - 68 - 69 74 viii DANH M C HÌNH Hình Hình - Cây FPHình - Cây FPHình - Cây FPHình Hình Hình - 14 18 19 19 20 22 prelarge large 1-itemset 26 Hình 41 Hình Hình Hình Hình Hình Hình Hình Hình Hình - 60 61 61 61 62 62 63 64 65 ix 5.3.2.3 Lu ng Hình 5-8 5và FUFPkhi minsup 0.05 minconf 0.5 Chúng tơi Hình - 64 -Growth s504002(s,"0 4") ==> s503001(s,"0 4") conv:(3.08) lift:(3.54) cosine:(0.58) lev:(0.07) s506001(s,"4 5") ==> s503001(s,"0 4") conv:(1.67) lift:(2.52) cosine:(0.36) lev:(0.03) s503002(s,"0 4") ==> s503001(s,"0 4") conv:(2.24) lift:(3.1) cosine:(0.41) lev:(0.04) s006004(s,"0 4") ==> s006001(s,"0 4") conv:(1.89) lift:(5.44) cosine:(0.53) lev:(0.04) s006001(s,"0 4") ==> s006004(s,"0 4") conv:(1.92) lift:(5.44) cosine:(0.53) lev:(0.04) s503001(s,"0 4") ^ s006018(s,"0 4") ==> s504002(s,"0 4") cosine:(0.51) lev:(0.04) conv:(2.34) lift:(4.94) s503001(s,"0 4") ^ s504002(s,"0 4") ==> s006018(s,"0 4") cosine:(0.4) lev:(0.04) conv:(1.76) lift:(3.03) s006018(s,"0 4") ^ s504002(s,"0 4") ==> s503001(s,"0 4") cosine:(0.46) lev:(0.04) conv:(4.67) lift:(3.98) Hình - s506001(s,"4 5") ==> s503001(s,"0 4") Giáo viên 65 5.3.2.4 L c lu t v ( minsup 0.05 minconf 0.5) : s503002(s,"0 4") ==> s503001(s,"0 4") conv:(2.24) lift:(3.1) cosine:(0.41) lev:(0.04) s006018(s,"0 4") ^ s504002(s,"0 4") ==> s503001(s,"0 4") cosine:(0.46) lev:(0.04) conv:(4.67) 66 lift:(3.98) NG K T 6.1 Nh ng công vi c hi n làm c - - - Growth Ch -Growth - Ngoài theo cơng trình [9] - tài 67 ng phát tri n ti p theo 68 Tài li u tham kh o [1] Maimon, O., Rokach, L (2010), The Data Mining and Knowledge Discovery Handbook 2nd edition, Springer [2] International Journal of Approximate Reasoning, 40 (1-2), pp 44-54 [3] Baker, R.S.J.d ( International Encyclopedia of Education (3rd edition), Vol 7, pp 112-118 [4] IJCA Proceedings on International Conference on Emerging Technology Trends (ICETT), (3), pp 20-25 [5] Expert Systems with Applications, 33(1), pp 135-146 [6] A IEEE Transactions on Systems, Man, and Cybernetics, 40(6), pp 601-618 [7] Heiner, C., Heffernan, Supplementary Proceedings of the 13th International Conference of Artificial Intelligence in Education, AIED 2007 [8] Romero, C., Ventura, S., Pechenizkiy, M., Baker, R.S (2011), Handbook of Educational Data Mining, Chapman and Hall/CRC, Data Mining and Knowledge Discovery Series [9] Merceron, A., Ya International Conference on Educational Data Mining, Montreal, Canada, pp 57-66 [10] Proceedings of Artificial Intelligence in Education (AIED2005), Amsterdam, Netherlands, IOS Press, pp 467 474 69 [11] Procedia Social and Behavioral Sciences, 2(2) , pp 5251 5259 [12] and solutions of applying association rules mining in learning Proceedings of the International Workshop on Applying Data Mining in e-learning (ADML'07), Crete, Greece 2007, pp 13-22 [13] , [14] of strong International Workshop on Applying Data Mining in e-learning (ADML'07), Crete, Greece 2007, pp 3-12 [15] Romero, C., Ventura, S., de Castro, C., Hall, W., Ng, M.H (2002), -based Adaptive Systems for Web-based Education, Malaga, Spain [16] Interestingness Measure for Association 8th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San francisco, USA, pp 67-76 [17] International journal of information theories and applications, 10(4), pp 370-376 [18] GESTS International Transactions on Computer Science and Engineering, 32(1), pp.71-82 [19] Agrawal, R., Imielin 70 The ACM SIGMOD conference, pp 207-216 [20] Proc AAAI Knowledge Discovery in Databases, pp 371-376 [21] Jiawei, H., Kamber, M (2006), Data Mining: Concepts and Technique 2nd edition, Morgan Kaufmann Publishers [22] Frequent Pattern Mining in the Incremental Database without Next-Generation Applied Intelligence, Lecture Notes in Computer Science, Vol.5579, pp 757-766 [23] of Discovered Association Rules in Large Databases: An Incremental 12th International Conference on Data Engineering, New Orleans, pp 106 114 [24] association rules based on adjusting FPDatabase Systems for Advanced Applications, Lecture Notes in Computer Science, Vol 2973, pp 417 424 [25] Seventh International Database Engineering and Applications Symposium, pp 111 116 [26] Fifth IEEE International Conference on Data Mining, pp.274 281 [27] García, E., Romero, C., Gea, M., Castro, C (2009) Proceedings of the 2nd International Conference On Educational Data Mining, pp 299-306 71 [28] ng for Proceedings of the 3rd International Conference On Educational Data Mining, pp 91-100 [29] -free association rule mining with Information Series, Vol RNTI-E-20, Hermann-Editions, pp 251 254 [30] GONZALEZ, A.Z, MENENDEZ, V.H., MÉNDEZ, M.E P., Proceedings of the 4th International Conference On Educational Data Mining, pp 321-322 [31] Evaluation Measures with Educational Datasets: A Framework to Proceedings of the 1st International Conference On Educational Data Mining, pp.177-181 [32] ElThe 2008 international Arab Conference of Information Technology (ACIT2008) Conference Proceedings [33] European Journal of Scientific Research, 47(1), pp 156-163 [34] Tair, M.M.A, ElInternational Journal of Information and Communication Technology Research, 2(2), pp 140-146 [35] ACM Computing Surveys, 38(3), Article [36] Han, J., Pei, J., Yin, Y The 2000 ACM SIGMOD international conference on management of data, pp 1-12 72 [37] mining algorithm using pre-large itemsets Intelligent Data Analysis, 5(2), pp 111 129 [38] Hong, T.P., Lin, Expert Systems with Applications, 34(4), pp 2424 2435 [39] Lin, C W., Hong, T P., Lu, The Pre-FUFP Algorithm Expert Systems with Applications, Vol.36, pp 9498-9505 [40] frequent itemsets based on the trie structure and the pre-large The 2011 IEEE International Conference on Granular Computing, pp 369 - 373 [41] IEEE Transactions on Knowledge and Data Engineering, 18(4), pp 472 481 [42] Bond 38(7-9), pp 739-751 [43] Ke, Y., Cheng, J., Ng, W (2008), An information-theoretic approach to quantitative association rule mining , Knowledge and Information Systems, Vol.16 n.2, pp.213-244 73 PH L C 0.03, minconf 0.05 cosine s006001(s,"0 4") : 56 ==> s006004(s,"0 4") : 44 conf:(0.79)lift:(5.86)cosine:(0.44)lev:(0.03)conv:(3.73) s006004(s,"0 4") ^ s501128(s,"0 4") : 71 ==> s501127(s,"0 4") : 41 conf:(0.58)lift:(6.94)cosine:(0.46)lev:(0.03)conv:(2.1) s006002(s,"0 4") : 163 ==> s006001(s,"0 4") : 90 conf:(0.55)lift:(2.99)cosine:(0.45)lev:(0.04)conv:(1.8) s006004(s,"0 4") ^ s501127(s,"0 4") : 75 ==> s501128(s,"0 4") : 41 conf:(0.55)lift:(6.95)cosine:(0.46)lev:(0.03)conv:(1.97) s006001(s,"0 4") : 246 ==> s006004(s,"0 4") : 131 conf:(0.53)lift:(2.19)cosine:(0.46)lev:(0.05)conv:(1.61) s501128(s,"0 4") : 105 ==> s501127(s,"0 4") : 53 conf:(0.5)lift:(6.07)cosine:(0.49)lev:(0.03)conv:(1.82) s006004(s,"0 4") ^ s504002(s,"0 4") : 50 ==> s503001(s,"0 4") : 41 conf:(0.82)lift:(8.61)cosine:(0.51)lev:(0.03)conv:(4.52) s006004(s,"0 4") ^ s503001(s,"0 4") : 57 ==> s504002(s,"0 4") : 41 conf:(0.72)lift:(9.89)cosine:(0.55)lev:(0.03)conv:(3.11) s504002(s,"0 4") : 97 ==> s503001(s,"0 4") : 68 conf:(0.7)lift:(7.36)cosine:(0.61)lev:(0.04)conv:(2.93) 74 s007002(s,"4 5") : 91 ==> s006002(s,"0 4") : 51 conf:(0.56)lift:(4.3)cosine:(0.41)lev:(0.03)conv:(1.93) s503001(s,"0 4") : 127 ==> s504002(s,"0 4") : 68 conf:(0.54)lift:(7.36)cosine:(0.61)lev:(0.04)conv:(1.96) s503001(s,"0 4") : 127 ==> s504002(s,"0 4") : 66 conf:(0.52)lift:(4.95)cosine:(0.49)lev:(0.04)conv:(1.83) s006018(s,"0 4") ^ s504002(s,"0 4") : 93 ==> s503001(s,"0 4") : 75 conf:(0.81)lift:(3.79)cosine:(0.46)lev:(0.04)conv:(3.85) s503001(s,"0 4") ^ s006018(s,"0 4") : 105 ==> s504002(s,"0 4") : 75 conf:(0.71)lift:(4.2)cosine:(0.49)lev:(0.04)conv:(2.81) s504002(s,"0 4") : 227 ==> s503001(s,"0 4") : 158 conf:(0.7)lift:(3.27)cosine:(0.62)lev:(0.08)conv:(2.55) s503001(s,"0 4") : 284 ==> s504002(s,"0 4") : 158 conf:(0.56)lift:(3.27)cosine:(0.62)lev:(0.08)conv:(1.86) s006001(s,"0 4") : 135 ==> s006004(s,"0 4") : 72 conf:(0.53)lift:(4.91)cosine:(0.51)lev:(0.04)conv:(1.88) s501128(s,"0 4") : 88 ==> s501127(s,"0 4") : 45 conf:(0.51)lift:(5.68)cosine:(0.44)lev:(0.03)conv:(1.82) 75 s006018(s,"0 4") ^ s504002(s,"0 4") : 83 ==> s503001(s,"0 4") : 70 conf:(0.84)lift:(3.98)cosine:(0.46)lev:(0.04)conv:(4.67) s504002(s,"0 4") : 172 ==> s503001(s,"0 4") : 129 conf:(0.75)lift:(3.54)cosine:(0.58)lev:(0.07)conv:(3.08) s503001(s,"0 4") ^ s501127(s,"0 4") : 64 ==> s504002(s,"0 4") : 47 conf:(0.73)lift:(5.7)cosine:(0.45)lev:(0.03)conv:(3.1) s503003(s,"0 4") : 82 ==> s006018(s,"0 4") : 58 conf:(0.71)lift:(3.95)cosine:(0.41)lev:(0.03)conv:(2.69) s503001(s,"0 4") ^ s503002(s,"0 4") : 71 ==> s504002(s,"0 4") : 47 conf:(0.66)lift:(5.13)cosine:(0.43)lev:(0.03)conv:(2.47) s503002(s,"0 4") : 108 ==> s503001(s,"0 4") : 71 conf:(0.66)lift:(3.1)cosine:(0.41)lev:(0.04)conv:(2.24) ea5(s,"0 4") : 68 ==> s504002(s,"0 4") : 44 conf:(0.65)lift:(5.02)cosine:(0.41)lev:(0.03)conv:(2.37) s503001(s,"0 4") ^ s006018(s,"0 4") : 110 ==> s504002(s,"0 4") : 70 conf:(0.64)lift:(4.94)cosine:(0.51)lev:(0.04)conv:(2.34) ea5(s,"0 4") : 68 ==> s503002(s,"0 4") : 42 conf:(0.62)lift:(7.63)cosine:(0.49)lev:(0.03)conv:(2.31) s006001(s,"0 4") : 130 ==> s006004(s,"0 4") : 70 conf:(0.54)lift:(5.44)cosine:(0.53)lev:(0.04)conv:(1.92) s501128(s,"0 4") : 80 ==> s006004(s,"0 4") : 43 conf:(0.54)lift:(5.43)cosine:(0.42)lev:(0.03)conv:(1.9) 76 s006004(s,"0 4") : 132 ==> s006001(s,"0 4") : 70 conf:(0.53)lift:(5.44)cosine:(0.53)lev:(0.04)conv:(1.89) s503002(s,"0 4") : 108 ==> s504002(s,"0 4") : 57 conf:(0.53)lift:(4.09)cosine:(0.42)lev:(0.03)conv:(1.81) s501128(s,"0 4") : 80 ==> s501127(s,"0 4") : 41 conf:(0.51)lift:(6.45)cosine:(0.45)lev:(0.03)conv:(1.84) 77 Ngày sinh: 2/7/1986 Email: thanbk2005@gmail.com 9/2005 tính 1/2010 9/2010 8/2013 Q TRÌNH CƠNG TÁC 6/2010 3/2011: trình viên Q Tr - Aptech, 392 4/2011- 186 78 ... LIÊN QUAN GT DL TT GC : Ghi TN 3.1 Khai phá lu t k t h p giáo d c Enrique García et al.[27] (recommendation learning - ientPredictive Apriori 31 sinh viên tham gia vào cosine (2) v Balcázar yacaree... Hình - Cây FPprefix (f, c, a) f, c, a a:2) b m b:1 - -4 18 Hình - Cây FP- Hình - Cây FP- 2.4 Khai phá lu t k t h 2.4.1 Thu t toán FUP Fast-UPdate algorithm) Cheung et al : 19 DB L Lk large k-itemset... I1I4 thêm vào L' 2.5: Cho CSDL ban vào ban c -6 ng - Giao d u TID Items 100 ACD 200 BCE 300 ABCE 400 ABE 500 ABE 600 ACD 700 BCDE 800 BCE ng - Giao d ch m i TID Items 900 ABCD 1000 DEF 24 ng - Large