_PROBLEM CoEPrA-2006_Regression_001 _GROUP_NAME Efstratios Georgopoulos _GROUP_MEMBERS Efstratios Georgopoulos, Technological Educational Institute of Kalamatas Assistant Professor and Hellenic Open University Computer Science Dept Tutor, sfg@teikal.gr George Zarogiannis, Hellenic Open University Computer Science Dept student, zarogiannis@gmail.com _ADDRESS Dr. Efstratios F. Georgopoulos, Technological Educational Institute of Kalamatas, 24100, Kalamata, Greece. tel.: +302721045278 _MODELING_PROCEDURE Proposed software model developed with Tree-Based Genetic Programming introduced by John R. Koza. This machine learning paradigm, uses genetic algorithm to evolve tree structures that represent computer programs. For an introduction, you may see www.genetic-programming.org. For performance optimization, steady state approach with tournament selection preferred against the use of generations. All original descriptors without any modification provided as input to the algorithm. Descriptors selected to appear in the final model by the overall evolution process. Twenty randomly selected records from calibration dataset used for validation and remaining records used for training. Best models from three different runs combined to create a final population that continued to evolve until the end of training. Finally, validation data used for a short time evolution, in order to select model with better generalization and further improve it. The final result used for prediction, scored 0.00924 Mean Squared Error in full calibration dataset. _PREDICTION Obj_00001 5.119 Obj_00002 6.073 Obj_00003 4.945 Obj_00004 5.800 Obj_00005 7.756 Obj_00006 5.983 Obj_00007 5.048 Obj_00008 4.429 Obj_00009 5.921 Obj_00010 5.227 Obj_00011 6.191 Obj_00012 3.179 Obj_00013 5.846 Obj_00014 4.884 Obj_00015 6.720 Obj_00016 5.977 Obj_00017 4.653 Obj_00018 6.411 Obj_00019 4.946 Obj_00020 4.816 Obj_00021 5.457 Obj_00022 5.219 Obj_00023 5.633 Obj_00024 5.781 Obj_00025 5.188 Obj_00026 6.403 Obj_00027 5.832 Obj_00028 5.829 Obj_00029 5.176 Obj_00030 5.460 Obj_00031 5.861 Obj_00032 5.842 Obj_00033 5.805 Obj_00034 6.122 Obj_00035 6.337 Obj_00036 1.178 Obj_00037 4.122 Obj_00038 5.900 Obj_00039 5.829 Obj_00040 5.094 Obj_00041 5.307 Obj_00042 5.335 Obj_00043 7.963 Obj_00044 5.328 Obj_00045 4.950 Obj_00046 5.935 Obj_00047 4.579 Obj_00048 6.303 Obj_00049 4.557 Obj_00050 4.562 Obj_00051 6.278 Obj_00052 5.896 Obj_00053 5.578 Obj_00054 4.429 Obj_00055 5.458 Obj_00056 5.989 Obj_00057 5.279 Obj_00058 4.685 Obj_00059 6.761 Obj_00060 4.487 Obj_00061 18.521 Obj_00062 5.040 Obj_00063 13.157 Obj_00064 4.461 Obj_00065 5.953 Obj_00066 9.134 Obj_00067 6.033 Obj_00068 6.437 Obj_00069 5.996 Obj_00070 5.492 Obj_00071 1.976 Obj_00072 5.352 Obj_00073 4.920 Obj_00074 5.449 Obj_00075 6.142 Obj_00076 4.809 Obj_00077 6.447 Obj_00078 6.456 Obj_00079 5.278 Obj_00080 6.218 Obj_00081 5.997 Obj_00082 5.476 Obj_00083 4.803 Obj_00084 5.671 Obj_00085 6.775 Obj_00086 6.166 Obj_00087 5.239 Obj_00088 3.487