Pattern classification with missing values using multitask learning
MetadatosMostrar el registro completo del ítem
Área de conocimientoTeoría de la Señal y las Comunicaciones
PatrocinadoresThis work will stimulate future works in many directions. Some of them are using different error functions (crossentropy error in discrete tasks, and sum-of-squares error in continuous tasks), adding an EM-model to probability density estimation into the proposed MTL scheme, setting the number of neurons in each subnetwork dynamically using constructive learning, an extensive comparison with other imputation methods, to use this procedure in regression problems, and extending the proposed method to different machines, e.g., Support Vector Machines (SVM).
EditorialInstitute of Electrical and Electronics Engineers ( IEEE )
Cita bibliográficaGARCÍA LAENCINA, Pedro J., SANCHO GÓMEZ, José Luis, FIGUEIRAS VIDAL, Anibal R. Pattern classification with missing values using multitask learning. En: International Joint Conference on Neural Networks (2006: Vancouver, BC, Canadá). 2006 IEEE World Congress on Computational Intelligence. 2006 International Joint Conference on Neural Networks. IJCNN´06, July 16-21, 2006. Vancouver, BC (Canadá). 2006. 3594-3601p.
Palabras claveAprendizaje multitarea
Clasificación de patrones
In many real-life applications it is important to know how to deal with missing data (incomplete feature vectors). The ability of handling missing data has become a fundamental requirement for pattern classification because inappropriate treatment of missing data may cause large errors or false results on classification. A novel effective neural network is proposed to handle missing values in incomplete patterns with Multitask Learning (MTL). In our approach, a MTL neural network learns in parallel the classification task and the different tasks associated to incomplete features. During the MTL process, missing values are estimated or imputed. Missing data imputation is guided and oriented by the classification task, i.e., imputed values are those that contribute to improve the learning. We prove the robustness of this MTL neural network for handling missing values in classification problems from UCI database.
El ítem tiene asociados los siguientes ficheros de licencia: