Automatic temporal framing detection for news analysis

Abstract This paper introduces an analytical framework for automatic temporal framing detection in news analysis, combining framing theory with computational techniques. Responding to the need for systematic tools to trace frame evolution in changing media contexts, it integrates text mining, network analysis, and dynamic community detection. Applied to Italian media coverage of the controversial Reddito di Cittadinanza policy, the framework reveals the relational and temporal dynamics of framing. The study highlights theoretical and practical implications, advancing the view of framing as a dynamic, context-sensitive process and demonstrating the potential of computational methods in large-scale media research. It offers a scalable, replicable tool that connects qualitative insights with quantitative precision, laying the groundwork for future studies across media environments.