Decision-tree algorithm provides one of the most popular methodologies for symbolic knowledge acquisition. The resulting knowledge, a symbolic decision tree along with a simple inference mechanism, has been praised for comprehensibility. The most comprehensible decision trees have been designed for perfect symbolic data. Over the years, additional methodologies have been investigated and proposed to deal with continuous or multi-valued data, and with missing or noisy features. Recently, with the growing popularity of fuzzy representation, some researchers have proposed to utilize fuzzy representation in decision trees to deal with similar situations. This paper presents a survey of current methods for Fuzzy Decision Tree (FDT) designment and the various existing issues. After considering potential advantages of FDT classifiers over traditional decision tree classifiers, we discuss the subjects of FDT including attribute selection criteria, inference for decision assignment and stopping criteria. To be best of our knowledge, this is the first overview of fuzzy decision tree classifier.

Background. Aberrant lipid metabolism is an alteration common to many types of cancer. Dysregulation of lipid metabolism is considered a major risk factor for bladder cancer. Accordingly, we focused on genes related to lipid metabolism and screened novel markers for predicting the prognosis of bladder cancer. Methods. RNA-seq data for bladder cancer were obtained from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases. The nonnegative matrix factorization (NMF) algorithm was used to classify the molecular subtypes. Weighted correlation network analysis (WGCNA) was applied to identify coexpressed genes, and least absolute shrinkage and selection operator (LASSO) multivariate Cox analysis was used to construct a prognostic risk model. External validation data and in vitro experiments were used to verify the results from in silico analysis. Results. Bladder cancer samples were grouped into two clusters based on the NMF algorithm. A total of 1467 genes involved in coexpression modules were identified in WGCNA. We finally established a 5-gene signature (TM4SF1, KCNK5, FASN, IMPDH1, and KCNJ15) that exhibited good stability across different datasets and was also an independent risk factor for prognosis. Furthermore, the predictive efficacy of our model was generally higher than the predictive efficacy of other published models. Distinct risk groups of patients also showed significantly different immune infiltration cell patterns and associations with clinical variables. Moreover, the 5 signature genes were verified in clinical samples by quantitative real-time polymerase chain reaction (qRT-PCR) and immunohistochemistry, which were in agreement with the in silico analysis. For in vitro experiments, knockdown of IMPDH1 markedly inhibited cell proliferation in bladder cancer. Conclusion. We established a 5-gene prognosis signature based on lipid metabolism in bladder cancer, which could be an effective prognostic indicator.