Abstract：Learning to Predict Citation-Based Impact Measures Citations implicitly encode a communitys judgment of a papers importance and thus provide a unique signal by which to study scientific impactEfforts in understanding and refining this signal are reflected in the probabilistic modeling of citation networks and the proliferation of citation-based impact measures such as Hirschs h-indexWhile these efforts focus on understanding the past and presentthey leave open the question of whether scientific impact can be predicted into the futureRecent work addressing this deficiency has employed linear and simple probabilistic modelswe show that these results can be handily outperformed by leveraging non-linear techniquesIn particularwe find that these AI methods can predict measures of scientific impact for papers and authorsnamely citation rates and h-indiceswith surprising accuracyeven 10 years into the futureMoreoverwe demonstrate how existing probabilistic models for paper citations can be extended to better incorporate refined prior knowledgeWhile predictions ofscientific impactshould be approached with healthy skepticismour results improve upon prior efforts and form a baseline against which future progress can be easily judgedRetrieval and IndexingIn this situationwhen the influence of the paper and the author is predicted and the system is indexedthese resources can be more effective in the library systemespecially when looking for some of them as some When high influence may have resources that may affect specific fieldsThis study uses machine learning methods to predict the scientific influenceespecially the influence of papers and authorssuch as reference rates and H indexStudy the use of a large number of feature projectsand then use the regression model to supervise learning to predict the impact