Learning belief networks from data is NP-hard in general. A common search method used in heuristic learning is the single-link lookahead. It cannot learn the underlying probabilistic model when the problem domain is pseudoindependent. In learning these models, to explicitly trade model goodness of fit to data and model complexity, parameterization of PI models is required. In this work, we present an improved result for computing the maximum number of parameters needed to specify a full PI model. We also present results on parameterization of a subclass of partial PI models. Keywords: probabilistic reasoning, knowledge discovery, data mining, machine learning, belief networks, model complexity.