Similarity between semantic spaces

One of the challenges in Latent Semantic Analysis (LSA) is to decide which corpus is best for a speci…c applica-tion. Furthermore, there are some parameters in LSA, such as the size of the corpus, the weight (local or global) functions, number of dimensions to keep, etc. that are important in generating high quality LSA spaces but dif-…cult to determine. In this paper, we provide a general method to measure similarity between semantic spaces. Using this method, one can evaluate semantic spaces (such as LSA spaces) that are generated from di¤erent sets of parameters or di¤erent corpora. The method we have developed is generic enough to evaluate other dif-ferent kinds of semantic spaces