Transforming Management with AI, Big-Data, and IoT

This chapter focuses on discovering the most significant disruptive technologies such as artificial intelligence (AI), the Internet of Things (IoT), and cloud in data storage. Since digitization has become an essential part of daily life, data collection has resulted in the growth of a vast amount of data that can be used in several valuable application areas. Hence, effective data storage is required in the form of clouds for analysis and utilization in various business and service areas, comprising smart health, smart agriculture, smart transportation, smart city, etc. Due to the explosion of IoT devices and the requirement to handle an exponentially increasing amount of data, businesses can look into the developments of cloud-based IoT solutions, and AI offers quick analysis, prediction, and services to many smart applications. Cloud storage is based on data storage in cloud that offers reliable, flexible, and cost-effective services to its customers via public, private, hybrid, and community-based clouds.

The current development and growth in the social and industrial sectors have evolved the use of sensing and smart devices, collectively called the Internet of Things (IoT). Huge volume of information being produced and collected by these IoT devices is also called Big Data, which needs to be managed, processed, and analyzed for the development of the social sector such as health care, education and smart community, and industrial sectors like manufacturing and production. Big Data analytics and processing have contributed to the advancement of society as well as improved the industrial processes. High speed and continuous sensing generate huge volumes of complex data, and processing and analyzing these data in a certain time limit is a big challenge. For a real-time decision-making and monitoring system, data processing and analysis are challenging due to limited computational, communicational, and storage resources. Big Data analytics and processing tools can be applied over these massive data as per the type of application and outcome required. Different types of applications need to be supported by a variety of tools based on different principles and approaches. Big Data is a set of voluminous and heterogeneous information in a coordinate, semi-structured, and unstructured form. Modern innovation in machine learning (ML), deep learning (DL), and artificial intelligence (AI) fulfills the requirement of Big Data analytics processing for advanced real-time decision-making, monitoring, and controlling systems. Classification and processing of stream of data generated by sensing devices help predict future insights. It identifies information required to control and monitor decisions for an individual, a society, an organization, and an industrial application. In this chapter, we discuss the algorithm, principal tools, and technologies required for Big Data analytics and processing over data gathered by different sensing and Internet of Things devices.

In any country or organization where an election takes place, the voting procedure plays the most vital role. It must be secure and transparent at the same point in time. The e-voting procedure adapted by many countries provides a secure solution, but whenever this much sensitive data comes in digital format, security and transparency must be the point of consideration. In this chapter, we are untangling several electronic voting issues and provide the solution for different issues by using blockchain technologies. The main concern for this chapter is to propose a secure and enhanced e-voting platform using decentralized blockchain technology so that the voting procedure can be transparent and immutable, which leads to a voting platform where the voter can have the transparency of their vote and at the same of time no one will be able to hack or manipulate the casted votes. The author also has performed and explained the comparison between the existing voting system and the proposed voting system.

Agriculture is an imperative occupation since the historical backdrop of humanity is kept up. Artificial intelligence (AI) is leading to a revolution in agricultural practices. This revolution has safeguarded the crops from being affected by distinct factors like climate changes, the porosity of the soil, etc. Artificial intelligence finds many applications in the agricultural sector, including crop monitoring, soil management, pest detection, and a lot more. The primary idea of using artificial intelligence in agriculture is due its adaptability, superiority, and cost-viability. The customary techniques that were utilized by the farmers were not adequate enough to satisfy their necessities. In this way, new computerized strategies were presented. These techniques fulfilled the food prerequisites. The primary focus of this chapter is to review the different utilizations of artificial intelligence in farming. These innovations save the abundant utilization of water, pesticides, and herbicides; they likewise help in the proficient utilization of labor and improve the quality.

Big Data is currently a buzz phrase in both the institutional and IT industry, with the phrase being used to represent a wide field of perception, varying from deriving data from external sources, collecting and managing it, such data with scientific methods instruments. In the digital and estimating nature, data is produced and stored at a pace that swiftly surpasses the threshold limit. When data is transferred and shared at optical fibre and broadcast channels at the speed of light, data aggregate and market growth rate will rise. However, such big data’s accelerated growth pace poses many hurdles, such as accelerated data increase, transfer speed, diverse data, and security. Numerous areas of science such as biology, astrology, physics, business, etc. are creating tremendous volumes of data in organized and non-organized data. They necessitate discovering out valuable erudition of the bulky, clamorous data.

This chapter covers the importance of Digital Marketing. Digital marketing has become a very important tool for all the business men because with the increased usage of the Internet many companies have started using this platform for promoting their products. Digital marketing uses some tools or methods such as Search Engine Optimization (SEO), Content Marketing, Search Engine Marketing (SEM), Artificial Intelligence, Programmatic Advertising, Chat Bots, Video Marketing, Social Messaging Apps, Visual Search and Social Media Stories. Digital marketing also gives a platform for customer’s feedback so that the company can change the requirements according to their needs and uses; this is why digital marketing, despite the challenges, is famous at present.

A rapid rise in the number of cars running on roads has increased the need for parking spaces due to which many difficulties are faced by drivers to find a parking space in car parking areas in big cities. Utilizing the capabilities of computer vision, the authors propose an automated parking space detection and management system. This chapter presents an approach for classifying parking lots either filled or empty based on the Convolutional Neural Network (CNN) running on the administrator’s or owner’s system. Our application provides an efficient and robust way of occupancy detection in various light conditions, weather conditions, presence of shadows, and partial occlusions. The authors used a publicly available CNRPark dataset that contains nearly 9000 images captured from various cameras in different conditions and with obstacles. Along with this, the LSTM model is trained based on the classification data provided by the CNN model collected over a long period to predict the future availability of parking lots based on the specific day of the week, festival, and national holiday. Additionally, a deep learning model is used as an alarming system to detect unwanted events occurring in the parking space. This security system which detects oddity helps to improve the reliability of the system and cameras. No manual monitoring is required due to this anomaly detection system which takes the same input as the slot detection model. All these features make our system more efficient, reliable, and cheaper than traditional methods of parking space management systems such as barriers, sensors, etc. This system is served to the user through a web app and Android app from where the user can gather useful information.

We live in a world fast moving towards cashless transactions and performing almost all transactions online. Online transactions are either conducted via UPI payments or plastic money, that is, credit cards. There has been an immense rise in use of credit cards, which has also resulted in increased frauds being carried out with the same. In this chapter, we aim to discuss a number of such frauds and means of handling these frauds. We have focused on classifying algorithms using Machine Learning and Artificial Intelligence (AI) to detect fraud and propose an extension to verification of user while performing transactions by facial verification. The chapter holds a number of classification techniques and methods to handle skewed and imbalanced data, followed by face recognition methodologies. Certain techniques exist that remove skewness before classifying, but these generally have poor rate of learning. Sampling methods are often used for handling imbalance in data. Several AI algorithms such as Artificial Immune Systems (AIS) can be used in combination with some basic classifier. Their full potential is yet to be extracted. The field of face recognition is moving at a fast pace. With face liveliness detection gaining grounds, it is very tough to forge it. Thus, a robust verification system can be built using it. The face recognition library has recently been added as a separate library in Python for the sole purpose of face detection and recognition. We would be working on it along with other techniques available to do the same.

With the rise of new technology, the Internet of Things (IoT) undertakes a dynamic and central role in healthcare, smart homes, retail analytics, agricultural machinery, etc. Cloud computing serves a platform which is used as a base technology for IoT by allowing data transfer, storage, and accessing applications through the Internet. Association of cloud computing with IoT can provide countless opportunities but there is also drawback of various security threats. IoT with cloud computing is a distribution system which is vulnerable to various malicious attacks. Machine learning is a powerful tool for analysing the data and predicting normal or abnormal behaviour of connected IoT devices. It is a subsidiary of artificial intelligence which focuses on creating applications that learn from big data and generate results without being explicitly programmed. It leads to the transformation in security of IoT systems from a manageable communication between the devices to secure IoT-based systems.

The job of software is to perform its intended functionality and to satisfy the customer. But it cannot happen all the time. What is the reason behind it? Why does not the functionality of the software satisfy the customer all the time? The reason behind is that the software’s functionality may be good for today but may be obsolete for tomorrow. This happens because as the time moves ahead, customer requirements change with respect to time, and currently working software may not work tomorrow. This chapter addresses this crucial issue and proposes an evolving agent architecture αβEA (αβEvolving Agent) which evolves itself to accommodate new changes that occur over a period of time in the functionality. A purchase agent has been developed using the proposed architecture for the experiment. Experiment on the proposed model shows that software developed using the proposed αβEA agent satisfies the customer for today as well as for tomorrow.

Nowadays, the usage of IoT (Internet of Things) devices is rising exponentially. It is predicted that the number of IoT devices being used will explode in the near future, and thus the data captured by these devices. This growing demand cannot be handled by existing technologies that incorporate network infrastructure comprising cloud services and big data analytics. Though we have large data centres with huge computational capacity and cloud architecture for gathering the data from IoT devices and making the data available to these remote devices, as the number of devices are growing along with their heterogeneous nature, the existing solution will no longer be enough to handle them. Also, as cloud and network architecture are going to change, so does the big data analytical part to cater the changing needs of users. Therefore, a new scalable, efficient yet cost-effective network infrastructure must cater to this ever-changing demand of users and increase IoT devices. The chapter introduces the concept of network architecture based on virtualization technique, thereby using SDN (Software-Defined Network) to embed the network elements on software rather than on specialized hardware and can be readily leased out from the available pool. In this chapter, we are also going to evaluate the infrastructure, limitations of SDN as well as cloud IoT. A new alternative framework is SDN-IoT investigated based on pros and cons to control congestion and traffic monitoring in the cloud architecture for supporting the distributed processing at edge nodes of the network, thereby involving edge, cloud computing, and big data analytics.

Malware has emerged as a major threat to computer systems the way the use of complex computer software is increasing nowadays, along with the security of that system which is also becoming a big concern. Malware is rapidly penetrating the security circle of computing devices. The trick to detect such malware is possible through machine learning, and it is also necessary to prepare computer resources that they can identify and combat the malware. Types of malware are increasing, and related threats are also increasing. This chapter covers good knowledge of machine learning to prevent malware attacks on computing devices and understand the important mechanisms of machine learning concerning malware attacks’ current trends.

This chapter explores options trading. Many models have been developed to help the market participants analyse the market situation and price movement of options. The first CLP scheme gives insight regarding the complex logic programming, which in conjunction with Black-Scholes model establishes the relationship among volatility, expiry time, underlying stock price, etc. Then, the agent-based model explores the CDA market. It discusses the agent-based model for strategy formulation. Various parameters such as delta, gamma also play an important role in these processes. Then Gaussian curve distribution gives an insight into price movements of options trades.

In recent times, the amplified applications of big data, artificial intelligence and IoT are used to explore valuable insights for decision-making. Recent developments in the computer platforms, sophistication in networking technologies and Information and Communication allow the adoption of IoT in a variety of applications. Though academic and practitioners worked in the domain in the past, there are many instances that warrants academic document that exposes comprehensive idea in this field. This chapter discusses various aspects related with artificial intelligence, big data, Internet of things (IoT), analytics, and sensing data to offer ideas in this domain. In addition, it includes IoT architecture layers and proposed model by combining big data, IoT, analytics and IoT data sensing sources. This proposed model aims to offer wide-ranging usage and application of big data, IoT, and IoT data sensing sources.

Internet of Things (IoT) has revolutionized different technological fields and applications. Many sensory devices are connected in the network, and an enormous amount of real-time data is associated with that in the form of audio, video and measured numerical data. Deep learning is an advanced field of machine learning that tries to make the computer understand from the input data, and data passes through multiple layers to make necessary prediction and classification. Higher-level features are extracted progressively from the raw input data from multilayer architectures. Integration of deep learning with IoT can efficiently manage these multidimensional large sensory data to extract information from the data, make new predictions, and have a more efficient control mechanism. Different aspects of IoT and deep learning are discussed in this chapter, with various applications in different fields in a comprehensive manner. Multiple applications and challenges are also discussed with current solutions and future directions.

Presently, the use of Twitter is increasing, and occurrences of large number of tweets are one of the important sources of personal thoughts and opinions. In social media, sentiment analysis is a significant type for analysis of text to make choice to find out the negative and positive thoughts of the users. In this chapter, we have analysed sentiment analysis of tweets using two machine learning models (Logistic Regression and Decision Tree) to identify the best machine learning algorithms for tweet data analysis. Further, data pre-processing (tokenization and stemming) and data visualization are performed. Data engineering principles are applied to measure the performances and improve the results. Data engineering displays the statistics with different labels, hash tags and word frequency tables. Finally, the performance of both the machine learning algorithms is evaluated using F-1 score. Results demonstrate 4% increase in model performance if Logistic Regression with a particular feature is used.

In the modern world, online crowdfunding plays an important role where many investors can fund projects presented by various creators. Our project aims to compare the pros and cons of conventional crowdfunding and blockchain crowdfunding. In conventional crowdfunding, one faces many issues such as transparency issues, fraudulent issues, investor abuse etc. However, to overcome these issues, blockchain crowdfunding comes into play. Blockchain crowdfunding helps to overcome the issues faced in conventional crowdfunding. To implement blockchain crowdfunding, we have proposed a model named ‘Block Funding’, which is made using Ethereum smart contracts. It consists of a web app made of React/Next.js and Ethereum smart contracts used in the backend. It primarily focuses on all the basic crowdfunding features as well as voting through blockchain. Moreover, the model is deployed on a Rinkeby test network.