| The main subject is the application of machine learning techniques to improve understanding in complex interstitial lung diseases, in particular sarcoidosis. We hypothesize that sarcoidosis is not a single disease but a collection of genetically complex diseases with a wide degree of clinical heterogeneity. Knowledge of the individual groups (diseases) may be crucial in providing patients with effective treatments, for the prognosis of the disease as well as for future scientific research. Therefore, our primary goal is to find a categorization of sarcoidosis by means of unsupervised learning techniques. In addition to identifying the diagnostic categories, this research studies the problem of predicting the correct diagnostic class based on genetic information, especially SNPs. Another purpose of the study is to search for the optimal diagnostic policy for sarcoidosis patients using machine learning algorithms. An attempt will be made to generalize the results to other interstitial lung diseases. |
