| Title |
Comparison of customers look-alike models / |
| Translation of Title |
Potencialių klientų modelių palyginimas. |
| Authors |
Kuzma, Pijus |
| Full Text |
|
| Pages |
46 |
| Keywords [eng] |
logistic regression ; neural networks ; random forests ; classification ; direct marketing |
| Abstract [eng] |
Methodology and its implementation for binary classification problem with continuous and binary independent variables is presented. Using this methodology, multiple classification algorithms can be implemented and their performance evaluated and compared to each other. Depending on the actual classification problem and appropriate evaluation criteria, the best performing model can be chosen that would optimally solve the classification problem. Also, by comparing the records of each model deciles, it allows seeing if the models are predicting the same records. |
| Dissertation Institution |
Kauno technologijos universitetas. |
| Type |
Master thesis |
| Language |
English |
| Publication date |
2015 |
