Sains Malaysiana 34(1): 93-99 (2005)

 

Sistem Kebal Buatan untuk

Pengecaman Digit

(Artificial Immune System for Digit Recognization)

 

 

Siti Maryam Shamsuddin, Anazida Zainal & Shahliza Abdul Halim

Fakulti Sains Komputer & Sistem Maklumat

Universiti Teknologi Malaysia

81310 Skudai, Johor, D.T.

 

 

ABSTRAK

 

Sistem Kebal Buatan (SKB) adalah satu bidang biologi yang telah membuka lembaran baru kepada penyelidik sains komputer untuk menggabungkan konsep kebal buatan di dalam penyelidikan yang berkaitan seperti pencerobohan sistem keselamatan. SKB tabi'i adalah suatu sistem pembelajaran adaptif iaitu yang mempunyai ciri selari dan mekanisma pelengkap bagi pertahanan terhadap unsur asing atau bakteria yang memasuki tubuh badan manusia. SKB bertindak secara tak linear dan menggarap konsep biologi seperti pengelasan terhadap sel kendiri dan sel tak kendiri. Rencana ini memperihalkan pelaksanaan kaedah SKB menggunakan pendekatan pilihan negatif bagi proses pengelasan dan pengecaman corak terhadap digit sifar hingga digit sembilan. Data bagi setiap digit tersebut disari menggunakan kaedah momen tak ubah dan diwakili sebagai rentetan 8 bit. Setiap kelompok digit dikelaskan sebagai kendiri bagi menghasilkan data tak kendiri atau pengesan. Ini bermakna terdapat 10 kelas pengesan yang dijana, dan proses padanan antara data kendiri dan tak kendiri dilaksanakan menggunakan operasi XOR. Penjanaan keputusan bagi pengelasan untuk suatu digit dihitung berasaskan kepada nilai peratusan yang terhasil, iaitu nilai yang dijana merupakan nilai yang digunakan bagi mengecam digit yang wujud pada data ujian. Semakin besar nilai peratusan yang diperolehi, maka semakin hampir nilai tersebut kepada digit yang hendak dicam.

 

Kata kunci: Sistem Kebal Buatan; Pengecaman Digit

 

 

ABSTRACT

 

Artificial Immune System (AIS) is an emerging field to the computer scientists and most of the recent works concerning AIS is in the area of Intrusion Detection System (IDS). AIS is based on human immune system. It is distributed in nature, deploys the adaptive learning and exercises complementary mechanism to defend human body from bacteria or foreign elements. Artificial Immune system is non-linear and adopts the biological concept in classifying self against the non-­self cells. This paper discusses on the implementation of AIS using the Negative Selection Algorithm in classifying and recognizing patterns on digits (0 to 9). Data for every digit is extracted using moment invariants and is represented in 8 bit string. There are 10 sets of detectors generated and the complementary process between self and non-self is done using the XOR operator. The result for classification for a digit is based on the percentage matched. Higher percentage indicates that the test data is closed to the digit to be recognized.

 

Keywords: Artificial Immune System; Digit recognization

 

 

RUJUKAN/REFERENCES

 

Alexander, T. & Dasgupta, D. 2000. A formal model of an artificial immune system. Biosystems 55: 151-158.

Anchor, K.P. Williams, P.D. Gunsch, G.H. & Lamont,, G.B. 2002.  The computer defense immune system: current and future research in intrusion detection. Proceeding of the 2002 Congress on Evolutionary Computation. Honolulu, HI, USA: 1027-1032.

Bellili, A. gilloux, M. & Gallinari, P. 2003. An MLP-SVM combination architucture for offline handwritten digit recognition. International Journal on Document Analysis and Recognition 5: 244-252.

Cheng-Lin, L. Nakashima, K., Sako, H. & Fujisawa, H. 2003.  Handwritten digit recognition: benchmarking og state-of-the-art techniques.  Pattern Recognition : The Journal of the pattern Recognition Society 36:  2271-2285.

De Castro, L.N. & Timmis, J.I. 2003. Artificial immune system as a novel soft computing paradigm.  Soft Computing Journal 7: 526-544.

Dipankar, D. & Nino, F. 2000.  A Comparison of Negative and Positive Selection Algorithms in Novel Pattern Detection. Proceedings of the IIEE International Conference on Systems, Man and Cybernetics (SMC), Nashville, October 8-11.

Duncan, R., De Castro, L.N. & Ayara, M. 2002. Negative Selection: How to Generate Detector. Proceeding of 1st International Conference on Artificial Immune System, University of Kent, Canterbury: 89-98.

Hu, M.K.1962. Visual Pattern Recognition by Moment Invariant. IRE Trans. Inform. Theory, IT-8 : 179-187.

Hunt, J. & Denise E. 1996. Learning Using An Artificial Immune System. Journal of Network and Computer Application 19:189-212.

Jerome, H.C. 2000. The Immune System As A Model For Pattern Recognition and Classification. Journal of American Medical Information Association 7(1): 28 - 41.

Kim, J. & Bentley, P.J. 2002. Towards an artificial immune system for network intrusion detection: an investigation of dynamic clonal selection. Proceeding of the 2002 Congress on Evolutionary Computation, Honolulu, HI, USA: 1015-1020.

Leedham, C.G. 1994. Historical Perspectives of Handwriting Recognition systems. IEEE Colloquimon on Handwriting and Pen-Based Inputs 1: 1-3.

Maslina Darus & Siti Mariyam Shamsuddin. 2001. Mencungkil Kesiratan Momen Geometri Takubah Terhadap Pengkelasan Fitur. Jurnal Teknologi Maklumat 13(2): 1-15.

Satyanath, S. & Sain, F. 2001. An AlS Approach to a Color Image Classification Problem in a Real Time Industrial Application. Proceeding of IEEE International Conference on Systems, Man, and Cybernetics, Tucson, AZ, USA, 2285-2290.

Steven A. & Stephanie, F. 1999. Immunity by Design: An Artificial Immune System. Proceeding of the Genetics and Evolutionary Computing Conference (CECCO), San Francisco, CA, 1289-1296.

Xiao, R.B., Wan, L. & Liu, Yong. 2002. A Framework of AIS Based Pattern Classification and Matching For Engineering Creative Design. Proceeding of the 1st International Conference on Machine Learning and Cybernatics, Beijing, 4-5 September, 1554-1558.

 

 

 

previous