Score Normalization Using Logistic Regression with Expected Parameters

摘要: State-of-the-art score normalization methods use generative models that rely on sometimes unrealistic assumptions. We propose a novel parameter estimation method for based logistic regression, using the expected parameters from past queries. Experiments Gov2 and CluewebA collection indicate our is consistently more precise in predicting number of relevant documents top-n ranks compared to state-of-the-art approach another estimate regression.