Benzetimli Tavlama Algoritması İle Eksik Veri Tamamlama

Öz

İstatiksel birçok yöntem eksik değerlere sahip veri setleri üzerinde çalışma kapasitesine sahip değildir. Bu nedenle, girdi olarak yalnızca tam veriyi kabul eden modellerin tahmin performansı önemli ölçüde düşmektedir. Eksik verilerin tamamlanması bunun için veri analizlerinde önemli bir yere sahiptir. Bu çalışmada kullanılan veri seti üzerinde eksik olan verilerin tamamlanma probleminin çözümünde sezgisel optimizasyon yöntemi olan Benzetimli Tavlama Algoritması(BTA) kullanılmıştır. Modern sezgisel teknikler, bir problem çözümünde, kendi yerel arama sistemleri ile en iyi sonuca ulaşmayı amaçlamaktadırlar. BTA performansını etkileyen en önemli değer başlangıç sıcaklık değeri (T0) olduğundan üç farklı sıcaklık değeri ile sonuçlar alınmıştır. To=100.000 değeri için %68, To=10.000 için %51 ve To=1.000 için %46’lik bir başarı elde edilmiştir

Anahtar Kelimeler

• Eksik veri
• Genetik Algoritma
• Benzetimli Tavlama Algoritması
• Sezgisel yöntemler

Öz

Many statistical methods are not capable of working on datasets with missing values. Therefore, the forecasting performance of models that accept only full data as inputs drops significantly. For this reason, completing missing data has an important place in data analysis. Simulated Annealing Algorithm (SAA), a heuristic optimization method, was used to solve the problem of completing the missing data on the data set used in this study. Modern heuristic techniques aim to achieve the best results with their local search systems when solving a problem. Since the most important value affecting SAA performance is the initial temperature value (T0), results have been obtained with three different temperature values. The following success rates were obtained: 68% for T0=100.000, 51% for T0=10.000 and 46% for T0=1.000.

Anahtar Kelimeler

• Missing data
• Genetic Algorithm
• Simulated Annealing Algorithm
• Heuristic Methods

Kaynakça

1. [1]. Sefidian A.M, Daneshpour N. Estimating missing data using novel correlation maximization based methods, Applied Soft Computing Journal, 2020; 91: 106249.
2. [2]. Rahman M.G, Islam M.Z. Missing value imputation using a fuzzy clustering-based EM approach, Knowl. Inf. Syst. 2016; 46 (2): 389–422.
3. [3]. Gopalakrishnan R, Guevara C.A, Akiva M. Combining multiple imputation and control function methods to deal with missing data and endogeneity in discrete-choice models, Transportation Research Part B, 2020; 142: 45–57.
4. [4]. Ye C, Wang H, Li J, Gao H, Cheng S. Crowdsourcing-Enhanced missing values imputation based on Bayesian network, in: International Conference on Database Systems for Advanced Applications, Springer; 2016: 67–81.
5. [5]. Mercaldo S.F, Blume J.D. Missing data and prediction: the pattern submodel, Biostatistics, 2020; 21(2): 236–252.
6. [6]. Zhiyong C, Longfei L, Ziyuan P. Yinhai Wang Graph Markov network for traffic forecasting with missing data, Transportation Research Part C; 2020.
7. [7]. Qin Y, Zhang S, Zhu X, Zhang J, Zhang C. POP algorithm: Kernel-based imputation to treat missing values in knowledge discovery from databases, Expert Syst. Appl., 2009; 36 (2): 2794–2804.
8. [8]. Molenberghs G, Thijs, H, Jansen I, Beunckens, C, Kenward, M.G, Mallinckrodt, C, Carroll, R.J. Analyzing Incomplete Longitudinal Clinical Trial Data; 2004.

9. [9]. Sayın A, Yandı A, Oyar E. Kayıp Veri ile Baş Etme Yöntemlerinin Madde Parametrelerine Etkisinin İncelenmesi, Journal of Measurement and Evaluation in Education and Psychology, 2017; 8(4): 490-510.
10. [10].Carpita M, Manisera M. On the imputation of missing data in surveys with Likert-type scales. Journal of Classification, 2011; 28(1): 93-112.
11. [11]. Demir E, Parlak B. Türkiye’de eğitim araştırmalarında kayıp veri sorunu. Eğitimde ve Psikolojide Ölçme ve Değerlendirme Dergisi, 2012; 3(1): 230-241.
12. [12]. Sezgin E, Çelik Y.Veri Madenciliğinde Kayıp Veriler İçin Kullanılan Yöntemlerin Karşılaştırılması, Akademik Bilişim Konferansı, Akdeniz Üniversitesi, 2013.
13. [13]. Little, R.J.A. , Rubin, D.B. Statistical Analysis with Missing Data: Second Edition. John Wiley and Sons, 2002
14. [14]. Şener Y. Veri Biliminde Eksik/Kayıp Verilere Yaklaşım Stratejileri ve Python (Pandas) Uygulaması, 2020
15. [15]. Min L, Yue C, Xiaojing S, Zhishan Z, Xiaoxiao Z, Xiuyu Z, Jun G. Simulated annealing-based optimal design of energy efficient ternary extractive dividing wall distillation process for separating benzeneisopropanol-water mixtures, , Chinese Journal of Chemical Engineering,2020.
16. [16]. Xiangzhen Z, Sanjiang L, Yuan F. Quantum Circuit Transformation Based on Simulated Annealing and Heuristic Search, , Transactions on Computer-Aided Design of Integrated Circuits and Systems, 2020.
17. [17]. Moriguchi K. Acceleration and enhancement of reliability of simulated annealing for optimizing thinning schedule of a forest stand, Computers and Electronics in Agriculture, 2020.
18. [18]. Kirkpatrick S, Gelatt C.D, Vecchi, M.P. Optimization by Simulated Annealing. Science new series, 1983; 220(4598): 671–680.
19. [19]. Songsheng T, Minjun P, Genglei X, Ge W, Cheng Z. Optimization design for supercritical carbon dioxide compressor based on simulated annealing algorithm, Annals of Nuclear Energy, 2020.
20. [20]. Asrul S.R, Ikram M, Mohd R. Mohd A.A. Energy Management Strategy of HEV based on Simulated Annealing, Int. J. of Integrated Engineering, 2020; 12(2): 30-37.
21. [21]. Lizhong Z, He M, Wei Q, Haiyan L. Protein structure optimization using improved simulated annealing algorithm on a three-dimensional AB off-lattice model, Computational Biology and Chemistry, 2020.
22. [22]. Jin C, Bin W. Flocking Control of Mobile Robots via Simulated Annealing Algorithm, Proceedings of the 39th Chinese Control Conference, 2020; 3931- 3935.
23. [23]. Tatsuya K, Hideharu K, Hiroyuki N, Tatsuhiro T. Using simulated annealing for locating array construction, Information and Software Technology 126, 2020.
24. [24]. Attiya I, Elaziz M, Xiong S. Job Scheduling in Cloud Computing Using a Modified Harris Hawks Optimization and Simulated Annealing Algorithm, Computational Intelligence and Neuroscience, 2020.
25. [25]. Hanine M, Benlahmar E H.A Load-Balancing Approach Using an Improved Simulated Annealing Algorithm, J Inf Process Syst, 2020;16:132-144.
26. [26]. Jafari H, Ehsanifar M, Sheykhan A. Finding Optimum Facility’s Layout by Developed Simulated Annealing Algorithm, Int. J. Res. Ind. Eng., 2020; 9(2): 172–182.
27. [27]. İlhan İ. A population based simulated annealing algorithm for capacitated vehicle routing problem, , Turk J Elec Eng & Comp Sci, 2020; 28: 1217–1235.
28. [29]. Xianze M, Yunpeng F, Junsheng Y. Estimating solubilities of ternary water-salt systems using simulated annealing algorithm based generalized regression neural network, Fluid Phase Equilibria, 2020.
29. [30]. Cunha M, Marques J. A New Multiobjective Simulated Annealing Algorithm—MOSA‐GR: Application to the Optimal Design of Water Distribution Networks, Water Resources Research, 2019.
30. [31]. Minghao G, Chunbo W, Baicheng L, Bin S. Yuanshen Huang Design and implementation of a Placido disk-based corneal topographer optical system based on aberration theory and simulated annealing algorithm,Optics Communications 475, 2020.
31. [32]. Liang F. Optimization Techniques Simulated Annealing A popular method for optimizing model parameters, 2020.
32. [33]. Tsai C.W, Hsia CH, Yang SJ, Liu SJ, Fang ZY. Optimizing hyperparameters of deep learning in predicting bus passengers based on simulated annealing, Applied Soft Computing Journal 88, 2020.