As social networks grow larger and larger, finding users of interest becomes an increasingly difficult task, making these networks, such as Twitter and Facebook, great scenarios for the application of recommender systems. In this work, we perform an extensive evaluation of content-based, collaborative-based, diffusion-based algorithms for user recommendation. We perform experiments on two real datasets from Twitter. In this work, a new user representation for tf-idf content-based algorithms is proposed. This representation captures the users’ interests more fully, by also taking into account the content posted by the people they follow. Our experiments show that this new representation outperforms the traditional one. We apply state-of-the-art collaborative filtering item recommendation algorithms in our user recommendation scenario. We also introduce and evaluate ProfileRank, a new diffusion-based model for measuring user influence and content relevance. ProfileRank is also applied to the task of user recommendation. Previous research has shown that there is value in combining different recommendation algorithms, as each algorithm has strengths and weaknesses. However, previous works have focused on specific classes of recommendation algorithms, or on naïvely combining different algorithms. In contrast, in this work we present a holistic hybrid framework that simultaneously takes into account content-based, collaborative-based, diffusion-based and user-based information. Our framework learns how to combine different sources of evidence (including the output from other algorithms) from the data itself, by using a Logistic Regression model. Therefore, instead of manually determining the importance of each source, or worse weighting all the sources equally, the appropriate emphasis given to each of the sources in our model comes from the data. Our experiments show that our algorithm outperforms current state-of-the-art algorithms for user recommendation.