woniu6667 / survival-time-prediction-of-a-patient-using-lstm-mlp-dbn

This repo is about predicting survival time of a patient using different deep learning techniques. Basically many models were developed before in literature .

Publication Details

Improving the accuracy of prediction of lung cancer patient survival time using LSTM Neural Networks.

International Conference of Sabaragamuwa University of SriLanka Nov 2019

Survival-Time-Prediction-of-a-patient-using-LSTM-MLP-DBN

This repo is about predicting survival time of a patient using different deep learning techniques. Basically many models were developed before in literature .

we tried to developed the system such that RMSE is minimum compared to works done before in literature.

Methods Used

1.MLP
2.DBM
3.LSTM

Abstract:

Outcomes for cancer patients have been previously estimated by applying various machine learning techniques to large data-sets such as the Surveillance, Epidemiology, and End Results (SEER) program database. In particular for lung cancer, it is not well under- stood which types of techniques would yield more predictive information, and which data attributes should be used in order to determine this information. In this study, a number of supervised learning techniques is applied to the SEER database to classify lung cancer patients in terms of survival, including linear regression, Decision Trees, Gradient Boosting Machines (GBM), Support Vector Machines (SVM), CNN and Deep belief model. Key data attributes in applying these methods include tumor grade, tumor size, gender, age, stage, and number of primaries, with the goal to enable comparison of predictive power between the various methods. The prediction is treated like a continuous target, rather than a classification into categories, as a first step towards improving survival prediction. We conclude that application of these supervised learning techniques to lung cancer data in the SEER database may be of use to estimate patient survival time with the ultimate goal to inform patient care decisions, and that the performance of these techniques with this particular data-set may be on par with that of classical methods.

Results

Models RMSE Standard Deviation Mean of Predictions Mean of residuals
LSTM 10.53 14.2652 42.8517 7.5264
MLP 1 14.8787 11.5504 45.4631 9.3820
MLP 2 14.9684 11.6146 46.3205 9.4452
DBN 16.399 7.4902 40.0 14.5900