In incomplete contexts, SE-ISI concepts contain abundant uncertain information, but not all SE-ISI concepts are necessary. This paper studies the theories and methods of SE-ISI concept reduction under different situations in incomplete contexts. First, SE-ISI concept reductions preserving positive information, generalized positive information and relation are defined, respectively. Then, the relationships between the three types of SE-ISI concept reduction are analyzed. Moreover, by introducing the SE-ISI representative concept matrix, the methods of obtaining three types of SE-ISI concept reduction are given. Finally, the characteristics and relationships of SE-ISI concepts under three types of SE-ISI concept reduction are discussed from the perspective of SE-ISI representative concept matrix.