5. Conclusion
As a conclusion, we think modelling is the key in big data analysis. Advanced and complex models can help to address difficult problems and improve performance. Statisticians have advantages to be a strong competitor. However, to achieve the desired goal, statisticians need to be better involved in other steps of the project including experiment design, data acquisition, cleaning and pre-processing (see other papers in this special issue, e.g. Smirnova et al., 2017). We need to pay more attention on how to extract information from big noisy data using summary statistics and to use simple models before we consider any complex and advanced methodology.
