Dynamic Time Warping is an algorithm that is used to compare two time series.
Suppose you say, " Let's all go for shopping and then have our lunch in a good restaurant."
You take 10 seconds to say this sentence. Your friend repeats this exact same sentence in 30 seconds.
Both sentences are same wit same words but one took 10 seconds to complete and other 30 seconds.
How to find that these are the same sentences? This was a challenge in speech recognition a few decades back.
Dynamic Time Warping algorithm solved the challenge by calculating the distance between the two time series.
Dynamic Time Warping shows to two time series to be almost similar by calculating the distance between the points in the two time series.
It finds the minimum path that minimizing the distance between any two points in the two time series.
We can use different distance measure like Euclidean, Manhattan and there are more distance measure that we can use.
Dynamic Time Warping lets us compare two time series and see if they have similarity or not.
I have written this blog post in which I explore the possibility of using Dynamic Time Warping in trading.
You can read this post I explore in detail whether we can develop a trading strategy using Dynamic Time Warping.
In trading, predicting the direction of the market is the most important if dynamic time warping can help, why not use it!
