kNN techniques are computationally efficient in the “prediction” phase, but take a long time to train.

The statement is actually the opposite. k-Nearest Neighbors (kNN) is a type of instance-based learning algorithm that is easy to implement. It doesn't require a training phase in the traditional sense. All the data is 'stored' instead of creating a model from which to make predictions. This is why it's often referred to as a 'lazy' learning algorithm.

However, this also means that kNN can be computationally expensive and slow during the prediction phase, especially when dealing with large datasets. This is because it has to calculate the distance of a point to every other point in the dataset in order to make a prediction. So, kNN is quick to train (as it simply stores the data) but potentially slow to predict.