Extracting the frequency from a complex signal is a common task in many applications, and multiple methods exist for accurate frequency estimation under different noise conditions. Iterative algorithms based on the interpolation of the discrete Fourier transform (DFT) are known to achieve high accuracy, and methods that employ auxiliary coefficients around the peak of the DFT are recurrent in the literature. This paper presents a novel iterative algorithm for frequency estimation based on successive parabolic interpolations of three DFT coefficients. Unlike other similar methods, which typically require auxiliary fine estimators for bias reduction, the proposed method refines the frequency estimate by progressively decreasing the offset of the DFT coefficients employed at each iteration. This approach eliminates the need for external correction steps and enhances estimation accuracy as the interpolation narrows around the true frequency. The algorithm achieves performance very close to the Cramér-Rao lower bound while maintaining computational efficiency, and the fine estimation step implemented can be flexibly applied to signals with or without zero-padding, making its use suitable for a wide range of signal processing applications. Simulations confirm the high accuracy and robustness to noise of the proposed estimator, showing comparable or better performance than existing iterative techniques.

The paper addresses the challenge of accelerating identification of changes in risk drivers in the insurance industry. Specifically, the work presents a method to identify significant news events ("signals") from batches of news data to inform Life & Health insurance decisions. Signals are defined as events that are relevant to a tracked risk driver, widely discussed in multiple news outlets, contain novel information and affect stakeholders. The method converts unstructured data (news articles) into a sequence of keywords by employing a linguistic knowledge graph-based model. Then, for each time window, the method forms a graph with extracted keywords as nodes and draws weighted edges based on keyword co-occurrences in articles. Lastly, events are derived in an unsupervised way as graph communities and scored for the requirements of a signal: relevance, novelty and virality. The methodology is illustrated for a Life & Health topic using news articles from Dow Jones DNA proprietary data set, and assessed against baselines on a publicly available news data set. The method is implemented as an analytics engine in Early Warning System deployed at Swiss Re for the last 1.5 years to extract relevant events from live news data. We present the system's architectural design in production and discuss its use and impact.