Artificial Neural Network (ANN) has been well recognized as an effective tool in medical science. Medical data is complex, and currently collected data does not flow in a standardized way. Data complexity makes the outcomes associated with intraoperative blood management difficult to assess. Reliable evidence is needed to selectively define areas that can be improved and establish standard protocols across healthcare service lines to substantiate best practice in blood product utilization. The ANN is able to provide this evidence using automatic learning techniques to mine the hidden information under the medical data and come to conclusions. Blood transfusions can be lifesaving and are used commonly in complex surgical cases. Blood transfusions come with associated risks and are costly. Anemia and clinical symptoms are currently used to determine whether a packed red blood cell transfusion is necessary. In this paper, we worked with unique datasets of intraoperative blood management collected from the electronic medical record of the Keck Medical Center of USC. We apply Multilayer Perceptron Neural Network to estimate missing values and predict the degree of post-operative anemia. Successful predictions of postoperative anemia may help inform medical practitioners whether there is a need for a further packed red blood cell transfusion.
[1]
M. Stone.
Cross‐Validatory Choice and Assessment of Statistical Predictions
,
1976
.
[2]
Vladimir N. Vapnik,et al.
The Nature of Statistical Learning Theory
,
2000,
Statistics for Engineering and Information Science.
[3]
Samy Bengio,et al.
Links between perceptrons, MLPs and SVMs
,
2004,
ICML.
[4]
Seung-Soo Han,et al.
Using neural network process models to perform PECVD silicon dioxide recipe synthesis via genetic algorithms
,
1997
.
[5]
Guang-Bin Huang,et al.
Learning capability and storage capacity of two-hidden-layer feedforward networks
,
2003,
IEEE Trans. Neural Networks.