ABSTRACT
In the last few decades, the expeditious expansion of the Internet and Internet of things (IoT) resulted in abrupt growth of data in almost every industry, business, and various fields of research. The embedded sensors, actuators, and software in IoT generate data in large volume and greater complexity. Analysis of these heterogeneous data with large volume and diverse dimensionalities is found to be challenging in extracting useful patterns from them and make it beneficial for the organization or society. Specialized tools are desired for analyzing and processing the huge chunk of data named “big data.” Regression and classification are the fundamental tasks for data mining, pattern recognition, and big data analysis in the field of data science. The major sources of data analytics and classification task for big data are generated through embedded sensors in IoT. Soft computing is another area of research where the feature sets of an entity are quantified as some degree of fuzziness rather than exactness. A decision tree (DT) is a rule-based, recursive, and top-down tree structure widely adopted for classification and regression. DT-based algorithms play a vital role in data mining, soft computing, and big data analytics that deal with multiobjective constraints. A number of applications based on DTs have been developed to address the IoT and big data research 100problems. Comparative analysis of various applications of DT can be very helpful in building models with better accuracy and precision. In this work, we perform an extensive state-of-the-art survey on the role of DT in the context of soft computing, big data, and IoT.
