This study introduces an interdisciplinary prediction framework as part of a novel approach that integrates the Inventive Design Method (IDM), Topic Modeling, and Generative AI to foster innovation across academic fields. Identifying interdisciplinary connections is essential for solving complex, multi-domain problems. Our research uses a supervised machine learning classifier to identify interdisciplinary documents within the Semantic Scholar corpus, extracting latent insights. The Text Convolutional Neural Network model performed best, achieving an F1 score of 0.80. We find that approximately 25% of human knowledge is interdisciplinary. This framework helps create comprehensive knowledge maps across multiple domains, promoting innovation through effective cross-domain knowledge transfer.