Data Mining Web-Tool Prototype Using Monte Carlo Simulations

Abstract: Facilitating the decision making process using models and patterns is viewed in this thesis to be really helpful. Data mining is one option to accomplish this task. Data mining algorithms can show all the relations within given data, find rules and create behavior patterns. In this thesis seven different types of data mining algorithms are employed. Monte Carlo is a statistical method that is used in the developed prototype to obtain random data and to simulate different scenarios. Monte Carlo methods are useful for modeling phenomena with significant uncertainty in the inputs. This thesis presents the steps followed during the development of a web-tool prototype that uses data mining techniques to assist decision-makers of port planning to make better forecasts using generated data from the Monte Carlo simulation. The prototype generates random port planning forecasts using Monte Carlo simulation. These forecasts are then evaluated with several data mining algorithms. Then decision-makers can evaluate the outcomes of the prototype (rules, decision tress and regressions) to be able to make better decisions.