| Περίληψη: | Economic problems' solving can be achieved by using many different methods. The most common is logistic regression. Just like many sciences, in economics, the evolution of technology has made possible, the solution of problems considered complex in the past and has contributed to the finding of new relationships that were hidden in the past. In this thesis we will try to analyze some of the problems that concern economics and now can be solved by using Data Mining algorithms. Our main topic will be Machine Learning, a broader sense of Data Mining, which is in fact, algorithms designed to ''learn'' through their use, finding relationships between variables and by solving a problem again and again, can become more efficient. We will accomplish a literature review and find for which sector of economics and which problem, the application of each algorithm is efficient. We must note that the efficiency of each algorithm is mainly based on the comparison with the logistic regression or sometimes by comparing different Machine Learning algorithms with each other. The results are very promising. Lastly, we conclude that in comparison with other sciences, in economics the use of these techniques is still in an early stage despite the urge of specialists for their further use. The proven efficiency of Machine Learning methods in the problems we introduce, should incite the economists to apply them in a greater degree.
|
|---|
