|
|
| |
| Use of Penalized Spline Linear to Identify Change in Pattern of Blood Sugar based on the Weight of Diabetes Patients |
Anna Islamiyati, Raupong, Anisa
|
Abstract:
Penalized linear spline regression contains first degree, knots and smoothing parameters which work simultaneously in the modeling. The flexible nature of the spline makes it possible to model data that has the possibility of variations of patterns in one model. That problem cannot be overcome by using linear regression in the parametric approach. Therefore, in this article, we show the use of the penalized spline linear regression models that contain several segments as a form of pattern in certain intervals. Furthermore, it is applied in the relationship of body weight with blood sugar in diabetic patients. Based on the model, we got 2 patterns of changes in blood sugar levels based on body weight. There is an upward pattern in body weight of less than 58.5 kg and a downward pattern in body weight from 58.5 kg upwards.
|
|
|
| | |