How Noisy and Missing Context Influences Predictions in a Practical Context-Aware Data Mining System

• Anca Avram ^[1] ; Oliviu Matei ^[1] ; Pintea, Camelia-M. ^[1] ; Pop, Petrica C. ^[1] ; Carmen Ana Anton ^[1]
  1. [1] Technical University of Cluj-Napoca

     Technical University of Cluj-Napoca

     Rumanía

     
• Localización: 15th International Conference on Soft Computing Models in Industrial and Environmental Applications (SOCO 2020): Burgos, Spain ; September 2020 / coord. por Álvaro Herrero Cosío, Carlos Cambra Baseca, Daniel Urda Muñoz, Javier Sedano Franco, Héctor Quintián Pardo, Emilio Santiago Corchado Rodríguez, 2021, ISBN 978-3-030-57802-2, págs. 22-32
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Resumen

  • The focus of this research is finding out how different levels of context noise and missing data, affect the overall prediction results in a Context-Aware Data Mining (CADM) system for predicting soil moisture. Experiments were performed using more machine learning algorithms and varying the levels of noise and missing context data in realistic scenarios. The results show that context with missing data has a higher impact on the predictions than noise. Results comparable to the clean context baseline are obtained when the 20% threshold of noise and missing data is not exceeded.