Association rule mining algorithms, such as the well-known Apriori algorithm, utilize support and confidence measures to mine interesting correlations in transactional data sets. Traditional support mechanisms cannot differentiate between projects of varying importance and therefore ignore uncommon but relevant terms. This paper introduces the concept of combining term frequency-inverse document frequency (TF-IDF) with weighted support, preferentially discovers association rules containing rare but key terms, and designs and optimizes the complex weighted Apriori algorithm. This study pioneers a novel TF-IDF weighted support metric that combines project support with the average TF-IDF of its component projects. This weighted support has been mathematically proven to comply with downward closure properties, enabling hierarchical exploration of weighted frequent patterns. This results in a more compact, high-quality rule set than traditional Apriori methods, while revealing dependencies ignored by confidence metrics.