The Role of IoT and Blockchain

Techniques and Applications

Sanjay K. Kuanar, Brojo Kishore Mishra, Sheng-Lung Peng, Daniel D. Dasig, Jr., Sanjay K. Kuanar, Brojo Kishore Mishra, Sheng-Lung Peng, Daniel D. Dasig, Jr.

• Editorial: Apple Academic Press
• Año: 2022
• ISBN:
  9781000071832
• Edición: 1
• Categoría: Computer Science
• Categoría: Databases

This volume provides informative chapters on the emerging issues, challenges, and new methods and state-of-the-art technologies on the Internet of Things and blockchain technology. It presents case studies and solutions that can be applied in the current business scenario, resolving challenges and providing solutions by integrating IoT with blockchain technology. The chapters discuss how the Internet of Things (IoT) represents a revolution of the Internet that can connect nearly all environment devices over the Internet to share data to create novel services and applications for improving quality of life. Although the centralized IoT system provides countless benefits, it raises several challenges. The volume presents IoT techniques and methodologies, blockchain techniques and methodologies, and case studies and applications for data mining algorithms, heart rate monitoring, climate prediction, disease prediction, security issues, automotive supply chains, voting prediction, forecasting particulate matter pollution, customer relationship management, and more.

Part I IoT Techniques and Methodologies

CHAPTER 1 Data Mining Algorithms for IoT: A Succinct Study

CHANDRAKANTA MAHANTY,¹ BROJO KISHORE MISHRA,¹ and RAGHVENDRA KUMAR²

¹Department of Computer Science and Engineering, GIET University, Gunupur, Odisha, India

²Department of Computer Science and Engineering, LNCT College, Jabalpur, India

ABSTRACT

IoT is a fresh idea which enables people to connect different sensors and intelligent machines to gather information from the surroundings in real time. It is anticipated that the number of effective IoT machines will increase to 0.01 trillion and 0.022 trillion by 2020 and 2025, respectively. Lots of analytical techniques are brought into IoT to render IoT smarter; data mining is among the most precious techniques. Data mining is the method of finding interesting knowledge and possibly relevant patterns from big data sets and using algorithms to extract crucial information. This article focuses on data mining framework for IoT, data mining functionalities and usage of data mining in IoT applications. After that, a survey on different data mining algorithms is presented. We also analyze the effectiveness and efficiency of different data mining algorithms (K-nearest neighbors, Naïve Bayes, support vector machine (SVM), C5.0, deep learning artificial neural networks (ANN), and ANNs). We reviewed the above mentioned algorithms and concluded that DLANNs, ANNs, C5.0 give relatively higher accuracy and memory-efficient as compared to other algorithms. To address IoT data mining problems such as managing large quantities of information, data analysis (DA), a big data mining system is suggested.

1.1 INTRODUCTION

Internet of things (IoT) relates to the next stage of the internet, containing hundreds of billions of nodes covering different elements from tiny ubiquitous sensor systems and portable equipments to big web servers and supercomputer clusters [1]. IoT operates with linking the entire world’s stuff via the internet [2]. IoT incorporates fresh computing and communications systems (wireless communications systems, radio frequency identification (RFID) technologies, actuators, and sensors) and develops the next generation of internet [3]. Several scientists who work in multiple areas such as scholars, institutes, and departments of government gave a lot of importance in changing the today’s internet by developing multiple technologies such as smart city, smart transportation, smart healthcare [4], smart home, and smart agriculture. IoT gets and stores a lot of information from sensors, wearables, smartphones, and other devices that are activated on the internet. In order to convert information into implementable information, it must be assessed utilizing suitable data mining techniques. Sensor information from a smart home, for example, is used for safety surveillance or home automation for elderly or disabled individuals, or traffic data is evaluated to determine an optimal ambulance path. The information gathered from IoT systems is being utilized for comprehend and supervise complicated environments within us, providing better decision-making, higher accuracy, effectiveness, and productivity. Analysis of the produced information must be recorded effectively in order to make this choice more precise and processing the same needs techniques. Data mining is among the most helpful techniques. Datasets store the information produced from IoT devices. The big volume of information processed using data mining technique estimates the model, generalizing it to fresh information as well. Data mining is therefore the method of finding precious data from bulk quantities of information stored in databases (DB) and data warehouses.

1.2 DATA MINING FOR IOT

Data mining in IoT is used to handle the big volume of information that IoT devices collect. Data mining includes discovering and analyzing information from the huge information set. Data mining’s primary aim is to explore helpful patterns from big knowledge set obtained from machines, sensors, IoT devices [5]. Discovery of knowledge, analysis of patterns and collection of data are the words used in the IoT for data mining. Data mining’s main goal is to create an effective and concise framework that is suited to the knowledge set. Thus, a number of researches focus on the use or development of efficient IoT data mining techniques. The findings outlined in [6] indicate that it is possible to utilize data mining algorithms to create IoT smarter which could provide better services. Based on prior information, a DM method could also be used to create a choice that let us tell the sale of a specific product, as example, cold drink sales higher in summer as compared to winter [7].

Data mining converts a set of data into a comprehensible framework and incorporates significant information that helps to gain insight into the raw data gathered from different IoT apps. IoT thus forms a network of devices that implanted with sensor, electronics, and network connectivity through which devices can gather and exchange information. The ideal combination of IoT and data mining leads to an innovative technology that benefits every part of the population. We were inspired to evaluate a data mining structure for IoT apps by the enormous quantity of data produced by IoT apps and knowledge discovery in databases (KDD). The IoT data mining framework is shown below by considering data mining for IoT.

Data mining is primarily divided into two procedures. First procedure is descriptive and the second procedure is predictive. Data mining is represented in a short and aggregated manner called descriptive data mining and it provides important overall information. In predictive information mining, information is evaluated in a series to build one or more information frameworks to estimate the actions of freshly created information sets [8] utilizing methods such as classification, clustering, regression, and trend analysis. In the course of knowledge discovery, data mining can be seen as an important technique. The stages which are required for this method depicted as below:

FIGURE 1.1 Data mining framework for IoT.

• Stage 1: Cleaning of Data: Noise and inconsistent information will be deleted in this process.
• Stage 2: Data Integration: It brings various sources of information.
• Stage 3: Selection of Data: In this level important information is recovered that is important to the method of evaluation.
• Stage 4: Data Transformation of Data: Aggregation activities is conducted in this stage in order to convert or enhance information into adequate knowledge for data extraction.
• Stage 5: Data Mining: Using smart techniques, the required information and patterns are obtained in this phase.
• • Stage 6: Evaluation of Patterns: Exploration of interesting patterns which depict expertise based on certain pleasant procedures.
• Stage 7: Knowledge Presentation: Using visualization and information depiction methods, information is provided to the customer in this phase (Figure 1.1).

1.2.1 DATA MINING FUNCTIONALITIES

1. Classification allocates objects to target types or classes in a set. Its objective is to forecast the target class correctly in the information for each situation.
2. Clustering identifies clusters of comparable information objects and forms a cluster group.
3. Association assessment is the identification of membership rules that exhibit circumstances of object-value that often happen in a specified set of information.
4. Time series assessment includes techniques and tools for the assessment of time series data to obtain reliable data and other information features.
5. Outer assessment defines and models conceptual frameworks or patterns for objects whose conduct evolv...