<p>Accurately predicting the consequences of rail transit accidents is critical for the effective formulation of emergency response plans and the mitigation of associated hazards. Owing to the complexity of rail transit systems, accident descriptions consist of multi-type data. As effective tools for information distortion, intuitionistic 2-tuple linguistic information(I2LI) is an important means to decision-making. In this paper, we propose multi-type data transformation model between I2LI and different data types, which can reduce the information loss effectively. To process data and discover knowledge, we introduce the notion of intuitionistic 2-tuple linguistic formal context (I2LFC) combing I2LI and formal context. Then, the intuitionistic 2-tuple linguistic concept lattice is constructed based on I2LFC, which can represent linguistic information more effectively and comprehensive in fuzzy linguistic information processing. In addition, intuitionistic 2-tuple linguistic decision rules are extracted using the finer relation of intuitionistic 2-tuple linguistic concept lattice, which are utilized to construct a intuitionistic 2-tuple linguistic rule base. Finally, an intuitionistic 2-tuple linguistic rule extraction algorithm is proposed for forecasting decision results, together with examples to illustrate the validity and rationality of the proposed algorithms.</p>