Narrow Results

Conjugate gradient methods constitute excellent neural network training methods which are characterized by their simplicity and their very low memory requirements. In this paper, we propose a new spectral conjugate gradient method which guarantees sufficient descent using any line search, avoiding thereby the usually inefficient restarts. Moreover, we establish the global convergence of our proposed method under some assumptions. Experimental results provide evidence that our proposed method is preferable and in general superior to the classical conjugate gradient methods in terms of efficiency and robustness.

Does a post with specific emotional content that is posted on Twitter by an influential user have the capability to affect and even alter the opinions of those who read it? Accordingly, “influential” users affected by this post can then affect their followers so that eventually a large number of users may change their opinions about the subject the aforementioned post was made on? Social Influence can be described as the power or even the ability of a person to yet influence the thoughts and actions of other users. So, User Influence stands as a value that depends on the interest of the followers (via replies, mentions, retweets, favorites). Our study focuses on identifying such phenomena on the Twitter graph of posts and on determining which users’ posts can trigger them. Furthermore, we analyze the Influence Metrics of all users taking part in specific discussions and verify the differences among them. Finally the percentage of Graph cover when the diffusion starts from the “influential” users, is measured and corresponding results are extracted. Hence, results show that the proposed implementations and methodology can assist in identifying “influential” users, that play a dominant role in information diffusion.