{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/time-series-prediction-for-graphs-in-kernel","title":"Time Series Prediction for Graphs in Kernel and Dissimilarity Spaces","arxiv_id":"1704.06498","date":"2017-04-21","proceeding":null,"authors":["Benjamin Paaßen","Christina Göpfert","Barbara Hammer"],"abstract":"Graph models are relevant in many fields, such as distributed computing,\nintelligent tutoring systems or social network analysis. In many cases, such\nmodels need to take changes in the graph structure into account, i.e. a varying\nnumber of nodes or edges. Predicting such changes within graphs can be expected\nto yield important insight with respect to the underlying dynamics, e.g. with\nrespect to user behaviour. However, predictive techniques in the past have\nalmost exclusively focused on single edges or nodes. In this contribution, we\nattempt to predict the future state of a graph as a whole. We propose to phrase\ntime series prediction as a regression problem and apply dissimilarity- or\nkernel-based regression techniques, such as 1-nearest neighbor, kernel\nregression and Gaussian process regression, which can be applied to graphs via\ngraph kernels. The output of the regression is a point embedded in a\npseudo-Euclidean space, which can be analyzed using subsequent dissimilarity-\nor kernel-based processing methods. We discuss strategies to speed up Gaussian\nProcesses regression from cubic to linear time and evaluate our approach on two\nwell-established theoretical models of graph evolution as well as two real data\nsets from the domain of intelligent tutoring systems. We find that simple\nregression methods, such as kernel regression, are sufficient to capture the\ndynamics in the theoretical models, but that Gaussian process regression\nsignificantly improves the prediction error for real-world data.","url_abs":"http://arxiv.org/abs/1704.06498v3","url_pdf":"http://arxiv.org/pdf/1704.06498v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"time-series-prediction-for-graphs-in-kernel","repo_url":"https://gitlab.com/bpaassen/graph-edit-networks","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"distributed-computing","task_name":"Distributed Computing"},{"task_slug":"gaussian-processes","task_name":"Gaussian Processes"},{"task_slug":"time-series-1","task_name":"Time Series"},{"task_slug":"time-series","task_name":"Time Series Analysis"},{"task_slug":"time-series-prediction","task_name":"Time Series Prediction"},{"task_slug":"regression-1","task_name":"regression"}],"methods":[{"method_slug":"gaussian-process","method_name":"Gaussian Process"},{"method_slug":"speed","method_name":"SPEED"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}