研究目的
To propose an inherently non-negative latent factor model for extracting non-negative latent factors from non-negative big sparse matrices efficiently and effectively.
研究成果
The proposed Inherently Non-Negative Latent Factor (INLF) model is highly efficient and effective on big sparse matrices, achieving high prediction accuracy with low computational burden. It inherently meets non-negativity constraints during the training process, making it suitable for industrial applications.
研究不足
The computational burden of the model is noted, particularly due to the expensive operation of raising the mathematical constant e to the power of an arbitrary real number, although this is mitigated through approximation techniques.
