Abstract:
This special problem aims to develop and evaluate of system for estimation of
cancer recurrence duration and evaluate users satisfaction for using the system.
There are 11 factors that may affect the cancer recurrence duration, i.e. age, gender,
occupation, hometown, primary site code, cancer recurrence code, histological code,
co-morbidity code, medical procedure, stage, and time of cancer recurrence. The
researcher selects data from 3 cancer groups, which always occur in Siriraj hospital,
i.e. 2,728 data of Breast cancer, 1,894 data of Lower gastrointestinal cancer and 8,356
data of Gynecologic and Cervix. Such data is collected from January, 2536 B.E. to
December, 2556 B.E. 3 classification algorithms, i.e. K-Nearest Neighbor (IBk), Decision
Tree and Naïve Bayes, used to compared the performances of the model. The best
algorithms used to develop the system, using System Development Life Cycle
(SDLC.) method. The result of system evaluation, by the experts, is at very good level
(Accuracy and Processing X = 4.6, S.D. = 0.5, Convenience and Security X = 4.5, S.D. = 0.4)
and users' satisfaction are at very good level (X = 4.7, S.D. = 0.5).