Computer Science · Artificial Intelligence
Combination of SOMs for Fast Missing Value
Antti Sorjamaa, Amaury Lendasse, Eric Severin
See on the mapAbstract
Abstract. This paper presents a methodology for missing value imputation. The methodology is based on a combination of Self-Organizing Maps (SOM), where combination is achieved by Nonnegative Least Squares algorithm. Instead of a need for validation as when using traditional SOMs, the combination proceeds straight into final model building. Therefore, the methodology has very low computational time. The combination of SOMs also increases accuracy at the same time. The performance is demonstrated using a database from corporate finance field. 1