Predicting potential lncRNA‒disease association pairs is an important issue in the field of biomedicine. Traditional lncRNA‒disease association prediction algorithms are mainly based on biological network models. For instance, RWRlncD is developed for predicting the potential lncRNA‒disease association by performing random walk with restart on lncRNA functionally similar networks (Sun et al., 2014). However, biological graph network algorithms often have certain limitations and low accuracy. With the development of machine learning, new ideas have been brought to the construction of the association prediction algorithm (Liu et al., 2016, 2019a, b; Wan et al., 2019). Recently simboost algorithm based on matrix decomposition is developed for predicting drug‒target and miRNA‒disease potential associations (He et al., 2017; Chen et al., 2019). Previous studies have shown that simboost can extract relevant features from known association pairs and use machine learning algorithms to evaluate the possibility of association pairs with unknown labels. We propose that this algorithm framework can be improved and applied to lncRNA‒disease association prediction.