MatrixH calculates the output of the hidden layer in aneural network.
1. Discussion and Experimental results (Zi?ba et al.,2014) presented Boosted SVM for discovering rules from data setof the post-operative survival expectancy in the lung cancer patientsand extracted 16 features forprediction the post-operative survival expectancy. We tries to provide a way for predicting the post-operative survivalexpectancy in the lung cancer patients with a thoracic surgery data set (Zi?ba et al.,2014), 470 sample and number of input variablesis 16 features that are defined below.f1. DGN: Diagnosis – specificcombination of ICDf2.
PRE4:Forced vital capacity – FVC f3. PRE5:Volume that has been exhaled at the end of the first second of forcedexpiration – FEV1 (numeric)f4. PRE6:Performance status – Zubrod scale f5.
PRE7:Pain before surgery f6. PRE8:Haemoptysis before surgery f7. PRE9:Dyspnoea before surgery f8.PRE10: Cough before surgery f9.PRE11: Weakness before surgery f10.PRE14: T in clinical TNM – size of the original tumourf11.PRE17: Type 2 DM – diabetes mellitus f12.PRE19: MI up to 6 months f13.
PRE25: PAD – peripheral arterial diseases f14. PRE30:Smoking f15.PRE32: Asthma f16. AGE:Age at surgery Output:Risk1Y: True value if died in 1 year survival period The number of patients whosurvive after surgery in comparison to patients who die is higher, in a oneyear period; from total samples,400 patients are positive (survival) and70 patients are negative (died).
It is important, selectingpatients for thoracic lung cancer surgery with a low risk of post-operative inshort-term 30-day period or long-term 1 or 5 year survival (Zi?ba et al., 2014). We consider 1 year survival period for prediction thepost-operative survival expectancy in the lung cancer patients in this paper.
We implementour proposed system in Matlab version 7.12 ona laptop,1.7 GHZ CPU, and we used the roots mean square error (RMSE) in order to determinethe evaluation indicator to determine the best method.