In the era of big data, leveraging advanced information retrieval (IR) techniques and knowledge-driven data models offers unprecedented opportunities to predict socioeconomic indicators accurately. This paper reviews the current state of knowledge-driven data models in information retrieval and their applications in forecasting socioeconomic trends. We examine various methodologies, from traditional statistical models to modern machine learning and deep learning approaches. By integrating domain knowledge into data models, we can enhance prediction accuracy and reliability. The review highlights key advancements, challenges, and future directions in this interdisciplinary field, emphasizing the importance of combining data-driven and knowledge-driven approaches for improved socioeconomic forecasting.