In the era of big data, the explosive growth of large-scale heterogeneous data provides a database for predictive analysis. The integration of intelligent IT algorithms and data processing technologies has become the main means to improve prediction accuracy and can be promoted in fields such as finance, business, and industry. The insufficient generalization ability of a single algorithm and the poor adaptability of the model make it difficult to implement and meet the demand for accurate prediction in complex scenarios. The structure and content of this article first summarize the current status of research on big data and intelligent algorithm prediction modeling, and clarify the existing research gaps. Secondly, it designs a technical solution for "data preprocessing intelligent algorithm fusion modeling model optimization verification" to improve data quality by improving data cleaning and feature construction methods; this article integrates deep learning into traditional statistical algorithms to construct hybrid prediction models, and introduces adaptive optimization strategies to optimize model parameters; Finally, empirical research is conducted on actual datasets from specific industries to validate the effectiveness of the model. Empirical evidence shows that the model proposed in this article is significantly better than traditional models in all indicators, with an accuracy improvement of 7 percentage points compared to random forests and a 14 percentage point improvement compared to logistic regression. In addition, it also provides greater scene adaptability and reliable technical support in complex scenarios to predict large-scale data.
Wang et al. (Thu,) studied this question.