A Multivariant Regression Model For Optimization Of Cassiterite Flotation

The planning of experiments and the importance of statistical interpretation of the results have been recognized for years by scientists and engineers. The Factorial Design of experiments is among the various methods used for this purpose and is being widely applied in all branches of research and production. The basic principles of this tool have been described in detail (1, 2, 3, 4) and several examples of its application in various fields have been published (5). In recent years the mineral industry has also used Factorial Design to improve the efficiency of research, testwork and plant operation (6, 7, 8, 9). This method permits for a qualitative determination of the influence of the variables studied in a process and also a qualitative measure of the interaction of the variables. By graphic means or derivative methods, one may also establish the optimum conditions of the process (ibid 7, 8, 9). It is desirable, however, to develop mathematical methods to evaluate quantitatively these effects and restrict the optimization to the practical limitations of the variables. It is also convenient to adopt a notation system to generalize the Factorial Design.