The rapid development of the wind industry in recent decades and the establishment of this technology as a mature and cost-competitive alternative have stressed the need for sophisticated maintenance and monitoring methods. Structural health monitoring has risen as a diagnosis strategy to detect damage or failures in wind turbine structures with the help of measuring sensors. The amount of data recorded by the structural health monitoring system can potentially be used to obtain knowledge about the condition and remaining lifetime of wind turbines. Machine learning techniques provide the opportunity to extract this information, thereby improving the reliability and cost-effectiveness of the wind industry as well. This paper demonstrates modeling damage equivalent loads of the fore-aft bending moments of a wind turbine tower with the advantage of using the neighborhood component analysis as a feature selection technique in comparison to common dimension reduction/feature selection techniques such as correlation analysis, stepwise regression or principal component analysis. For this study a one-year measuring period of data was gathered, pre-processed, and filtered by different operational modes, namely stand still, full load, and partial load. Finally, a sensitivity analysis was performed in the partial load model to determine the required length of the data collection campaign that guarantees the most precise results. The results indicate that applying neighborhood component analysis yields more conservative models regarding the number of features and equally accurate outcomes than traditional feature selection techniques.