SVRC MODEL, PREDICTING THE EVAPORATION HEAT OF A PURE ORGANIC COMPOUND, CAPABLE OF REFLECTING A NON-LINEAR FUNCTIONAL RELATION OF AN INDEPENDENT VARIABLE AND A SUBORDINATE VARIABLE

摘要

PURPOSE: A SVRC(Scaled Variable Reduced Coordinate) model, predicting the evaporation heat of a pure organic compound, is provided to reflect a non-linear functional relation of an independent variable and a subordinate variable, which an MLRM(Multiple Linear Regression Model) is not able to reflect, thereby supplying high prediction performance.;CONSTITUTION: Experimental data of hydrocarbon series among sample compounds is inputted. A molecular descriptor for the evaporation heat of a hydrocarbon series compound is prepared. A necessary parameter for a SVRC formula is obtained. A QSPR(Quantitative Structure-Property Relationship) model about the evaporation heat at a 0.55 times temperature of a normal boiling point is built. Prediction performance is tested on the experimental data. If the test is satisfied, an evaporation heat prediction value by the model is selected as an evaporation heat value.;COPYRIGHT KIPO 2012