| Tekijä: | Wilson, J. |
| Otsikko: | Integer programming formulations of statistical classification problems |
| Lehti: | Omega
1996 : DEC, VOL. 24:6, p. 681-688 |
| Asiasana: | MATHEMATICAL PROGRAMMING STATISTICAL METHODS CLASSIFICATION AND CODING |
| Kieli: | eng |
| Tiivistelmä: | This paper presents formulations of statistical classification problems using integer programming, attempting to maximize the number of observations that can be classified and utilize single-, multiple-, and hierarchical multiple function approaches to the problems. In all these approaches data and modelling sensitivity is apparent for any linear programming based approach and the modeller can't use techniques in a "black box" way. |
SCIMA