In this paper, we propose a new classification method that improves the support vector machines technique (SVM). It consists of the real time SVM (RTSVM) that uses an incremental version of SVM which is the LASVM. It also takes into account of new data over time. Actually, current classification techniques suffer from scalability problem. There is a permanent growing and evolution of data. Besides, there is a need of important memory capacity and execution time to deal with data stream. Although the improvement made to SVM to reduce the memory use and computational time in training phase, the obtained model in training phase cannot be applied to new observations in test phase without using the hole data. To overcome this issue and improve classification task in test phase, the RTSVM adapts the initial model produced by the LASVM. After that, the RTSVM updates and improves it in test phase by only using new data for re-training. As a result, our proposal considerably reduces the execution time and improves the accuracy especially in test phase. Empirical study shows RTSVM to be effective when using real-world datasets.
This paper intends to propose an on-line monitor ing system based on the incremental support vector machines (LASVM). In fact, the current monitoring system in ICU presents a real threat for the patient life due to the high rate of false or non significant alarms. In this paper we aim to improve the current system by applying an intelligent and on-line classification method (the LASVM). This method adds new instances of medical parameters of patients over time and deals with large amount of data streams in ICU. Besides, the LASVM generates an optimal model of prediction which provides a better and correct description of the different patients' states over time. All obtained results of the LASVM on real-medical databases prove the performance of this new system. Our proposal reduces the false alarms and conserves the high level of sensitivity compared to the standard SVM and the current system. I. I Databases CS-TA SVM-TA LASVM-TA II Expert I
Purpose: Choosing the relevant features is important to provide a better understanding of the data and improve the prediction performance. Thus, the main aim of this paper is to identify the risk factors of breast cancer.
Methods: focusing on two different datasets: Breast Cancer Surveillance Consortium (BCSC) and Breast Cancer Coimbra (BCC), we perform a comparative study of various feature selection methods: Filter Methods, Wrapper Methods and Embedded Methods. In addition, this work investigates the stability of these techniques when perturbation on datasets is added. Artficial Neural Network, Random Forest, SVM, Logistic Regression and Decision Tree are used for classification. Results: The results are compared when using all the features and when using only the top ranked. The classification performances are comparable in either cases. Furthermore, we found that invasive, glucose, resistin, insulin, leptin, age, adiponectin, BMI and HOMA are the most relevant features that promote breast cancer.
Conclusion: Our findings demonstrate that the identified feature selection methods can efficiently determine the risk factors of breast cancer.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.