Note that batch processing provides rigorous results
since more data can be used and it performs better
training of predictive models. But it is not feasible for
domains which need low-response time. Real-time processing
generally ensures low response time. However,
low-response time can be achieved at the expense of
less rigorous analysis of data. The hybrid approach is,
therefore, required so that application domains (using
Big Data) can benefit from both batch and real-time
processing. To obtain desired results under this approach,
both batch and real-time results are queried. The results
are then merged together, synchronized or composed.
Data acquisition and analysis become more complicated
under the approach [6].