Instructions
IoT, particularly in manufacturing, has opened up an entirely new set of data
collection methods that are providing the fastest-paced amount of data we have
seen to date. In this final culminating assignment, you will use the information
you have learned from the past seven weeks to produce a report on the use of
data mining techniques and IoT.
First, identify a company that is using IoT as some integral part of their business.
Then, identify data that are collected and ways in which the company acts on the
information. Identify elements that have been automated as well as those that
still require manual input by an employee. Identify the business goal and
additional information that the company may find useful to collect as it moves
forward with its IoT initiative. Document the hardware and personnel that are
needed to complete work in this space and make assumptions about the capacity
to describe potential needs in the future. Identify at least two types of logic
components or assumptions that must be built into the process, and at least two
statistical techniques that can be used to determine whether the IoT process is
resulting in the expected return on investment. Be sure you source and
thoroughly describe each component.
Finally, research data mining techniques and corresponding theories that relate to
the use of IoT and how this new trend will impact data mining research in the
future.
Support your assignment by citing the studies you analyzed. In addition to these
specified resources, other appropriate scholarly resources, including older
articles, may be included to support your report and any factual assertions you
make.
Length: 12-15 pages
References: A minimum of 6 scholarly sources
Your essay should demonstrate thoughtful consideration of the ideas and
concepts presented in the course and provide new thoughts and insights relating
directly to this topic. Your response should reflect scholarly writing and
current APA edition 7 standards.
ABSTRACT
In today’s world, it is practically impossible to connect without using the Internet. The Internet of things will have a significant effect on humans; this is because it will cause enormous beyond human understanding. The devices of IoT, in a substantial way, generate big data that is highly accurate, and they are valuable and, at the same time, useful. It becomes hard to excerpt the needed data or information from a customary of extensive statistics revealed by any given device. In ensuring that this tenacity is met, data mining helps accomplish the set mission.
Furthermore, in the smart system’s construction, data mining plays a significant role in ensuring that the intelligent system is in place. The intelligent system is essential in providing convenient services. It is needed in data extraction and information from the things that are connected. Several data excavating methods are employed. The procedures of different calibers are applied in helping mine the data; some of the algorithms used include association rule mining, clustering, and classification. This paper will look deep into the IoT and information mining and some of the mining methods, trials, and excavating matters with IoT.
Keywords: Frequent pattern, classification, clustering, Internet of things, and Data mining.
Introduction
Today many companies have embraced the idea of IoT. In this regard, this paper focuses on the Silverstein Properties, a real estate development company located in New York, and how it uses IoT in its operations. IoT is a critical component in the company and plays an outsized role in locating and communicating with the clients securely. The primary data that the company collects is mainly on the location of the properties, the information of the people in the features, and how they respond to the resources. With this kind of information, they can easily trace the uptake of their properties in the market, which is a very significant feature in business management. The main elements include smart homes; this is mainly concerted in making sure that the clients acquire excellent properties that meet all their needs. Apart from that, a security system ensures that their properties are well protected from intruders. IoT enabled security systems to attract clients as they are assured of 24/7 security in the premises(Tan et al., 2019).
IoT is widely used in helping connect everything in the world through the use of the Internet. Many applications have been made possible with the help of good progress in the communication of computers and information technology. Most people in the world are embracing technology, and this will ensure that, in the future, the world will be connected through the Internet. The next advanced generation will majorly be based on the IoT. It is stated that in the coming years, the world will be connected through millions of nodes objects with an extensive web server and cluster of a supercomputer. IoT, also in a significant way, plays a role in ensuring integration in the new technologies of computing and even communication.
In recent years, mobile devices’ introduction has made communication easy by ensuring different people in the world are connected through an interface. These have enabled accessible communication between people in various places in the world. Through mobile devices, people can easily connect everywhere in the world in the comfort of their homes. With the help of these kinds of devices, there are no restraints that corrupt the connection between different people. Researchers working with government institutions, academics, and institutes have been on a lookout for ways of modifying the Internet by coming up with various devices and designing systems in healthcare, smart home, global supply chain, fashionable pen, and intelligent transportation.
Business goal
The main business goal for Silverstein Properties Company is to ensure that they increase the revenue through IoT, to reduce the cost of operation, and also to mitigate business risk.
Revenue increase
The company is integrating its market automation platform with IoT data analytics to bring about better customer experiences. The customers who are using their services are tracked through the load signature; thus, they can quickly note the usage pattern. Pairing demographic information with the IoT usage data helps the company segment their users and target them with specific promotions based on their profiles and habits (Éloi & Basel, 2016).
Reducing the cost of operation
Reducing the cost of operation is the primary goal of many companies. The management and technical team work collaboratively to ensure low operating costs. Although it is a big challenge, it is the primary driver of most enterprise IoT projects. The chief operating expense is energy in most companies; it is essential to reduce the power consumption rate. IoT data analysis may suggest that the company is running from 7:30 a.m. to 4:30 p.m., but beyond that time, the lights may be left open, adjusting the lighting time could help a lot in saving power consumption rate.
Mitigate Business Risk
Silverstein Properties injects large sums in repair and maintenance. Sometimes the repair is costly, and in most cases, it is the cause of 42 percent of unplanned downtime. Therefore, the company needs to ensure that they run a good maintenance program and have a predictive method of when the maintenance is necessary to help plan for the services before it happens.
Data mining future trends
Multimedia data mining
These are considered to be the latest development methods that capture essential data perfectly. Different multimedia sources are used in the extraction of data. Some of the multimedia sources that are employed include images, audio, video, hypertext, and text. It includes the conversion of data into a representation that is in numerical and in a different format. These methods can majorly be used in classifications and clustering, performing similarity checks, and association’s identification.
Ubiquitous Data mining
This method mainly involves data mining from mobile devices to get the full individual information. Although these trends face enormous challenges, such as cost, privacy, and complexity, it brings many opportunities to be enormous in different industries and mostly in studying the interaction between humans and computers.
Distributed Data mining
This type of data mining is becoming more popular as it involves information mining of an enormous amount of data that is mainly stored in the company’s different locations or at various organizations. From the different areas, highly sophisticated algorithms are extracted, and thus it gives proper reports and insights based upon them.
Geographic and Spatial data mining
This new data mining involves extracting information from geographical data, astronomical, and environmental data, including images taken from outer space. This sort of data mining may bring to different light aspects such as topology and distance, which is majorly used in geographic systems information and other navigation applications.
Sequence and Time series data mining
It is mainly involved in the study of seasonal and cyclical trends. These are primarily essential for random events analysis, mostly outside the regular series of events. The retail companies mostly employ these methods to get to know the purchasing behavior of the customers.
Additional information that the company needs to collect for IoT initiatives
IoT initiatives require skills such as machine learning and AI, engineering, and security infrastructure. Employing professionals with background knowledge in python and java- will, in a significant way, help a lot in the IoT field, thus making the company employ IoT-integrated. The technical team in the company needs to have some programming languages like, for instance, python and java. If you are a novice, it is necessary to hone your programming language skills before taking another step in IoT developing solutions. Get to know the crucial role of sensors—proven expertise in nodes. The platform of the node is mainly used in building the IoT applications.
IoT hardware
IoT hardware mainly includes devices, such as sensors, bridges, and routing. These equipment manage vital tasks and functions such as communication, system activation, action specifications, and security. Some examples of sensors include cameras, light sensors, and motion detectors. All these hardware require qualified and competent personnel; for instance, the sensors need someone good in the installation process and control. Installers play a significant role in making sure that the devices are functioning properly.
Even though IoT systems are automatic and self-regulation, there needs for companies to employ technicians to check whether the sensors and other IoT devices are functioning well. For the IoT hardware devices, the number of personnel needed should be approximately six, two taking care of installation, two for the maintenance of the IoT hardware, and the other two who should be in the control room and ensure that things are running the right way.
Logic components
A logical system assigned by an administrative entity mainly in the entire landscape system is referred to as logic components. These involve mostly the project phases and the system roles; the system role is primarily concerned with system development.
Quality assurance system
The data mining with IoT in these cases is mainly concerned with testing and knowing if the data mining devices are working correctly. These are very important in ensuring that there is no room for blunders.
Production system
These are mainly concerned with ensuring that data mining with IoT is bearing fruits and bringing about a solution to the Internet’s future. These provide that words are made easy through the IoT devices that are used.
Statistical technique
Regression analysis
It is important when it comes to the identification of variables that have an impact on the topic of interest. The regression analysis mostly gives a chance to determine the factors that matter the most confidentially. It is also essential to note the kind of factors that can be ignored. For instance, there are critical factors to be considered in IoT, such as user experiences and security issues. In contrast, others may be overlooked or dealt with slowly like, for example, cloud computing. These can be dealt with as time goes, and it can be valuable in the future. The regression analysis is essential in determining how the factors influence one another. It helps the company realize a return on investment because it focuses on only things that matter at that particular moment.
Factor analysis
It is mainly concerned with data reduction; it is essential to ensure that data mining in IoT that is not needed is reduced to minimize the cost and increase its revenue. It aims at reducing individual items into fewer dimension numbers. They are mostly employed in the reduction of variables numbers in the regression models. The factor analysis in a significant way will help the company realize returns on investment, considering that its primary aim is to reduce the data while trying to increase its revenue.
Data from IoT
It is evident that with IoT, the data created may constitute information that is very useful to people. In recent times there have been some challenges and technical issues, and the research nowadays aims to ensure that the problems are handled in different methods on this kind of data. The question of Large IOT data can easily be solved by proving that a sensor is designed in such a manner that they can collect only data that are useful and interesting. In a significant way, these will help lower the input data complexity in the recent research trend. The most common handling data methods include feature selections, cloud computing, and distributed computing. It is estimated that with big data, there will be more number of patterns from applications and services. It is also an essential task to know the hidden information and ensures we know how to handle the big obtainable data using IoT.
Sensors are essential in setting up useful systems for smartphones or smart cities, devices. During this extensive data analysis process, it is possible to come up with many applications. The Knowledge Discovery in Databases is a very successful method used in fulfilling the role of finding the information hidden from the big data.
Following the steps of applied knowledge Discovery in Databases has the power of coming up with something or a pattern that is interesting from IoT. The steps discussed include gathering data, preprocessing the data, data mining, and analyzing decisions. The above data-mining step is fundamental and plays a significant role in getting exciting patterns or knowledge from the big data. Some of the issues that substantially impact the system’s performance and the quality service of IoT include data decentralized, data transmission, and data fusion.
Data mining for IoT
Data mining basics with IoT
When it comes to creating and analyzing the data, it becomes more comfortable when using data mining for IoT (Alessandro, 2016). For a long time now, there have been studies conducted in trying to solve the issue of coming up with big data on IoT without making use of efficient and effective analyzing tools. Where big data is used, it becomes challenging to apply most traditional algorithms and knowledge discovery in databases directly on a large amount of IoT data.
It is also a hard task in coming up with a data module of high performance of knowledge discovery in databases for IoT. The only solution will be to use the knowledge discovery in the database process, including mining algorithms, characteristics of data, and objective.
Objective
It is essential to set it clear and distinguish between the problems of the problems to get solutions, limitations, measurements, and assumptions. In a significant way, the above-described parameters will help a lot in making the objective of the problem clear.
Data
In data mining, user data characteristics play a vital role, and some of the features include size, representation, and distribution. There is a different method in processing the various data. For instance, Rj and Ri may not be similar or maybe similar, but they need to be inversely analyzed if the semantics of the data is various (Han & Kamber, 2012).
Mining Algorithm
According to the particularized needs of data application and objective, it is easy to design a data mining algorithm. The big question is, which is the best data in the mining algorithm to helps increase the system’s higher performance or get the best service in various IoT environments. The extraction of big data from IoT uses different techniques, and these techniques are named classifications, clustering, and association mining rule.
Classification
These are essential in helping categories in both the unlabeled and the labeled patterns. The labeled data mainly consider the evidence unlabeled with class and related tag to it with some information. A picture of an animal will be deemed untagged; on the other hand, if the same animal’s image has names or tags on it, for instance, the animal’s sound, it is considered to be labeled. The data that is labeled make judgments about an available piece of unlabeled data.
Clustering
These are mainly on classifying the unlabeled patterns. Some of the examples of trends that are unlabeled include human-created facts and natural samples. In addition to the above, unlabeled designs may also include audio recording, photos, news articles, videos, tweets, and x-rays. The unlabeled pattern has not specific knowledge about its constituent but mainly contains information on the data.
Association
It is mainly used in finding events from the pattern input that does not occur in a particular order. Association mining rule contains the sequential pattern and is used primarily in finding events from pattern input order.
There may be differences in finding the objective of hidden patterns depending on the goal. Many researchers make out different steps in trying to give better services by coming up with different techniques used in mining. System designing and overall study are desirable since a solitary technology or procedure will not work to extract useful information and make a decision.
The above figure clearly shows the possible combination of the technologies used in mining and extracting hidden information. The clustering algorithm is the first combination, and then it is followed closely by the algorithm of classification. The combination that follows involves the procedure in classification to use first after used; the clustering algorithm is then smeared to the data. Without knowing the pattern or input data, the first combined with a set of classifiers is created. Classification is the main base in the additional combination, and it is accountable for creating classifiers and set. The possibilities of the information that enters the IoT incrementally or pattern handling enable the combination of different mining technologies—like, for example, recognizing the behavior and the faces of humans not previously in the database knowledge.
Data mining with IoT challenges
The main challenges are the crime detection application to the next step, which mainly includes an advanced feature that averts crime. Another problem comes in during extraction of extensive data that is available in large storage data and to requires to be detected unreliably without identifying any noise in the large dataset. It is also a challenge when it comes to incomplete data and mining uncertain. It is becoming too complicated in the process of modifying the algorithm; it is not an easy task for providing security solutions for data sharing (Pal & Balamuralidhar, 2017). It is not also an easy task for converting IoT to generate data into knowledge that can be used to provide a convenient atmosphere for the user.
Handling and analyzing broad data is a tiresome task in the process of data mining. Coming up with an idea and ensuring that it is implemented to be efficient and better technology in mapping up with the other kinds of techniques is not easy. It requires a lot of time and knowledge to be able to join the two to be compatible.
It is also not possible to come up with a simple algorithm by building an intelligent system; this is because, in the process, there must be a fusion of several algorithms to a single algorithm. Internet is used in rapid speed devices, there may be an issue of connection that may also come along, and thus it becomes a problem without the Internet’s context.
It is also a hard task, and it is not easy building up an industrial intelligent IoT device, including green energy generations and the smart city. Unexpected use of consumer data and extensive data collection is also a primary challenge in security. There also arise security issues with the sharing of infrastructure and standards.
Major issues in data mining with IoT
The parallel programming design needs to be aligned so that there can be a straightforward application of every algorithm into it. There should also be the designation of the framework that is mainly concerned with security like data size growth, data sharing, and privacy. From the perspective of infrastructure, it is clear that IoT gives a high throughput, and on the other hand, it gives a low computation. Still, the algorithm of mining is designed in such a way that its device power consumption is low and also of small size. These, in return, results in issues in the infrastructure.
From the data’s perspective, there is redundant data created by the different sources by data gathering. For the system’s performance to be better, the unnecessary information needs to be filtered by the user. In addition to that, there may also be obstacles resulting from data generated from different sources. From the algorithm’s perspective, adding classifiers is needed in some IoT, while some require the addition of classifiers statistically. There is a need for the fusion of some mining technologies that can classify the classifier in a standard way.
Security and privacy will remain an issue because all the technologies and algorithms cannot outdo security and privacy problems. A good example is that most companies find it difficult to get data from different customers using various sources or devices. They can also use the data mining technique to come up with the kind of information they need to help increase sales volumes. Still, most of the issues are that almost every client will not readily disclose their privacy and security; these are things like their shopping behavior and retails.
Massive scaling is an open issue using a more significant number of data, name, protect, identify, maintain, and authenticate. Dependency and architecture, it is hard to develop an architecture that can easily connect many things to the Internet. There is dependence on many things such as robustness; if the clock drifts, the device’s location needs to be known, and the device location may not be that accurate. Therefore the chances of removing even one of them may cause an error
Conclusion
The idea of IoT has been born from automating, exploring the different available devices, and managing all the available instruments and sources from all over the world. Through the continuous association between humans and machines, there comes the need for data mining techniques. These data mining methods mainly support the optimization of any application and decision-making process. When joined together with IoT, data mining concentrates mostly on discovering the new patterns and, at the same time, useful from the massive data. The information is hidden and, at the same time, is very important and extracted using different algorithms (Aboul, 2020).
The techniques used for data mining mainly include pattern mining, classification, and clustering. In adding more literature on data mining and the introduction of IoT, the paper also looks at the open issues and challenges in data mining. Data mining algorithms, significant data characteristics, and the analysis of the problems are also described. At the initial stage of the development of IoT, most of the focus is on developing efficient preprocessing mechanisms and effectively expanding the technologies used in mining to clear the IoT data rules.
Besides trends of the research that is going on in the world nowadays are also discussed in the paper. The main research trends are data mining from the control of the application on devices and mainly the multimedia devices. Regulating energy is also a topic of study, such as uploading data to the Internet when devices are connected but shows there is a problem that IoT faces during data fusion and summarization of the data and abstraction of the data.
References
Alessandro Bassi, Bauer, M., Fiedler, M., Thorsten Kramp, Rob Van Kranenburg, Lange, S., Meissner, S., & Springer-Verlag Gmbh. (2016). Enabling Things to Talk Designing IoT solutions with the IoT Architectural Reference Model. Berlin Springer Berlin Springer.
Aboul Ella Hassanien, Ashraf Darwish, & Hesham El-Askary. (2020). Machine learning and data mining in aerospace technology. Cham Springer.
Éloi Bossé, & Basel Solaiman. (2016). Information fusion and analytics for big data and IoT. Artech House.
Han, J., & Kamber, M. (2012). Data mining : concepts and techniques. Elsevier.
Pal, A., & Balamuralidhar Purushothaman. (2017). IoT technical challenges and solutions. Artech House.
Tan, P.-N., Steinbach, M., Anuj Karpatne, & Kumar, V. (2019). Introduction to data mining. Pearson Education, Inc.
References
Alessandro Bassi, Bauer, M., Fiedler, M., Thorsten Kramp, Rob Van Kranenburg, Lange, S., Meissner, S., & Springer-Verlag Gmbh. (2016). Enabling Things to Talk Designing IoT solutions with the IoT Architectural Reference Model. Berlin Springer Berlin Springer.
Aboul Ella Hassanien, Ashraf Darwish, & Hesham El-Askary. (2020). Machine learning and data mining in aerospace technology. Cham Springer.
Éloi Bossé, & Basel Solaiman. (2016). Information fusion and analytics for big data and IoT. Artech House.
Han, J., & Kamber, M. (2012). Data mining : concepts and techniques. Elsevier.
Pal, A., & Balamuralidhar Purushothaman. (2017). IoT technical challenges and solutions. Artech House.
Tan, P.-N., Steinbach, M., Anuj Karpatne, & Kumar, V. (2019). Introduction to data mining. Pearson Education, Inc.
