The application of artificial intelligence on big data generated by smart cities: A review of the preprocessing phase

Abstract

As populations grow in size, new challenges arise in the big metropolitan cities. Resources need to be better distributed, traffic needs to be managed better, pollution needs to be kept at a minimum, etc. In order to facilitate the solutions for these kinds of problems cities are investing in new technologies with the aim to increase their sustainability and to improve the level of comfort for its citizens. This kind of cities are called “Smart Cities”.

Through these new information and communication technology solutions, a lot of data is being generated at high speeds every day. We can classify this huge amount of data as big data. Data in smart cities comes from all kinds of sources in all kinds of formats. In most cases the data is not ready to be used for analysis when obtained, it can contain some noise or incorrect values. For cities to be able to analyze the data and to extract knowledge from the data, the data first needs to be preprocessed. This is a crucial step in the analysis process.

In this paper, we first define the concepts of “Smart Cities” and “Big Data”. Thereafter, we will focus on the preprocessing steps in the analysis process. We explain what machine learning entails and what its dimensions are. Next, we will list the opportunities and challenges smart cities face and in the last section, we will present some preprocessing methods proposed in the literature to help with the challenges mentioned earlier.