Distributed recursive least-squares strategies over adaptive networks