| ITI əməkdaşlarının elmi işləri - məqalə |
| Biblioqrafik təsvir | |
| Mamedova , M.H. Selecting a machine learning algorithm for creating a hepatocellular carcinoma prediction system / M.H. Mamedova , Z.G. Jabrayilova , L. Garayeva // Proceedings of Azerbaijan high Technical Educational Institutions. - 2022. - N: volume 22, Issue 11 .- P. 116-129. | |
| Annotasiya | |
| The formation of e-health has stimulated the development of intelligent systems that provide information support for medical decisions. These systems are used to make a diagnosis, choose a more effective treatment method, predict, search for suitable conditions (precedents), control and schedule therapy, recognize and interpret images, monitor the clinical-pharmacological properties (toxicity) of drugs, etc. The basis of these systems are diseases in the specific subject area of medicine, their possible causes, development period, clinical manifestations, observed signs, symptoms, etc. Successes achieved in the creation of such systems that prevent errors in making medical decisions, have necessitated the creation of a system for diagnosis and prediction of hepatocellular carcinoma (HCC), known as liver cancer, which is the third leading cause of lethal cancer in the world. HCC is characterized by a set of clinical manifestations of critical conditions, each of which, in turn, is defined by a set of clinical signs and data. The analysis of these data shows that in the conditions of information abundance, the physician has to make a decision by referring to a part of the obtained information. As a result, errors occur in physicians’ decisions determined by certain combinations of a vast number of indicators and clinical signs. To predict HCC based on such numerous, diverse and heterogeneous unstructured data, preference is given to the method of artificial intelligence, i.e., machine learning. Machine learning enables data processing, presentation, and making predictions based on the results obtained. This article explores the possibility of applying machine learning algorithms to create an HCC prediction system and solves the problem of selecting the best algorithm. The study conducted in this regard uses the HCC Dataset database taken from Kaggle platform, and 49 features/attribute data of 165 patients are referenced. The experiment conducted for the prediction of HCC is presented in stages. In the first stage, pre-processing (or the process of database cleaning) is performed in order to bring the data into a uniform form, while the cleaning of unrelated and scattered data is implemented with the direct involvement of the user. Libraries Scikit-learn, Pandas, NumPy, etc. are used in the Jupiter programming environment to make the database beneficial and simplify data processing. Correlation heatmaps are used to find both linear and non-linear relationships between variables. Min-max scaling is applied to one or more feature columns to normalize the data. The second step analyzes all features present in the database, determines their types and target class. | |
| Elektron variant | |
| Elektron variant |
________
© ict.az http://ict.az/az