Some people say this can be done using Fibonacci levels. I have seen good charts in which price respects the Fibonacci levels or RSI or MACD. But of course, these people only show the evidence which confirms their argument and scrap the examples which are negative.
Personally, I don't believe its possible to consistently identify the end of a correction: you only know this when the correction has already ended because price has resumed the main trend. But not knowing the precise price when the correction ends should not stop you entering using whatever technical analysis evidence you can apply. As long as you have a stop to get you out for those occasions when statistics undermine your good analysis.
