This paper presents an application of kernel methods for extracting relations from unstructured natural language sources. The authors introduce kernels defined over shallow parse representations of text and design efficient algorithms for computing these kernels. They use Support Vector Machine (SVM) and Voted Perceptron learning algorithms to extract person-affiliation and organization-location relations from text. The experimental results show that the proposed methods outperform feature-based learning algorithms, with particularly promising results for the organization-location relation due to its more complex nature. The paper also discusses the importance of shallow parsing in information extraction and the advantages of kernel methods in handling long-range dependencies in text.This paper presents an application of kernel methods for extracting relations from unstructured natural language sources. The authors introduce kernels defined over shallow parse representations of text and design efficient algorithms for computing these kernels. They use Support Vector Machine (SVM) and Voted Perceptron learning algorithms to extract person-affiliation and organization-location relations from text. The experimental results show that the proposed methods outperform feature-based learning algorithms, with particularly promising results for the organization-location relation due to its more complex nature. The paper also discusses the importance of shallow parsing in information extraction and the advantages of kernel methods in handling long-range dependencies in text.