Linear Discriminant Analysis is a dimensionality reduction technique which reduces dimensions based on a supervised algorithm. The algorithm reduces number of features/variables based on predefined class/labels present in the dataset. It selects the derived features in such a manner that the classes present in data are easily distinguishable.
In this notebook, LDA is applied on iris dataset https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data
In addition to this, for any classification problem, PCA, LDA and then a classification technique can be used in the given order for good results.