A uniform sampling rate is when a time series is sampled at fixed intervals, say once a second. The majority of time series anaysis techniques assume a uniform sampling rate. However, there are many knds of data where this is not the case, for example, medical data, such as blood results, only occur when a patient visits a doctor or while in hospital the doctir decides a further test is needed; thus they are thus both sporadic and almost cedrtainly at times that are not typical of times when no sample is taken. Even where the intention is to have a uniform sampling rate, sensor failures or lost data may mean there are missing values.
Used on page 318