By comparing standard deviation from every other rating's time. By comparing account voting patterns across videos. By comparing registration time and voting ti

By comparing standard deviation from every other rating's time. By comparing account voting patterns across videos. By comparing registration time and voting time. By identifying grouped clustered votes time relationship. This just on top of my head, going to bet there is going to be more to be found in aggregate.