Locality sensitivity discriminant analysis-based feature ranking of human emotion actions recognition.

摘要: [Purpose] Computational intelligence similar to pattern recognition is frequently confronted with high-dimensional data. Therefore, the reduction of dimensionality critical make manifold features amenable. Procedures that are analytically or computationally manageable in smaller amounts data and low-dimensional space can become important produce a better classification performance. [Methods] Thus, we proposed two stage techniques. Feature selection-based ranking using information gain (IG) Chi-square (Chisq) used identify best selected for emotion different actions including knocking, throwing, lifting. Then, feature reduction-based locality sensitivity discriminant analysis (LSDA) principal component (PCA) transform space. Two-stage selection-reduction methods such as IG-PCA, IG-LSDA, Chisq-PCA, Chisq-LSDA proposed. [Results] The result confirms applying combined dimensional-reduction method increases performance classifiers. [Conclusion] dimension was performed LSDA by denoting highest importance determined IG Chisq not only improve effectiveness but also reduce computational time.