In the introduction part the overall view of this chapter is going to be discussed. Over the year’s Artificial intelligence and machine learning has been advancing by the professionals. Deep learning has reached a position where it has an impact not only on the academic level and on the industrial research but it also has an impact on the popular media channels. In this report all the points, facts, aspects, results, discussions and many things are going to happen on the development of the machine learning models. Researching of deep learning methods have got some excellent results after using deep learning models and they built AI systems that have an impact on various fields such as computer, videographer, video games, processing of natural languages and so on.
This research is going to be on the development of machine learning and deep learning models using new technologies. Researchers have understood that it is very much needed to offer a better understanding of AI systems, especially those which are based on machine learning (Ziogas, et al, 2021). It aims to establish strong foundations for the field of work. In this paper the authors have developed an algorithm using Python to develop the machine learning models. The algorithm is going to make changes on the data set of the existing models. Python can be used to determine the changes in the models as python is the main language of machine learning. As python is easy to learn and the syntax of python is easy, for machine learning python is used. Since python is portable and extensible, python can be used to perform the tasks of machine learning. Due to the adaptability of the python, the researchers and the data scientists have used python to perform those cross language tasks and to train the machine learning models. Since, python provides a lot of review of the codes and tools for testing, data scientists have gone for python to test and run the models.
1.3 Aims and Objectives
The main aim of this project is to develop the models of machine learning using python as the software. The authors have determined new methodologies for machine learning and to use it for real life applications.
To read specific literature reviews on the same topic in machine learning
To identify several significant chances of developments and be able to code the new algorithm
To identify the gaps of the literature and to create a new algorithm
To propose new methodology and to test and run the new data sets and use it as a real life application to test the efficiency of the new algorithm.
1.4 Research Questions
Q1: Why Python is needed to run this software?
Q2: What will be the efficiency of the new algorithm?
Q3: What aspects of the new algorithm are going to be there?
Q4: How is it related to real life applications?
Q5: What will be the length of the new data sets?
1.5 Research Hypothesis
H0: Reading specific literature reviews for too many times might be confusing and this research is going to be primarily analyzed.
H1: Too many types of ideas can be taken from different literature reviews so that a specific algorithm can be designed with so many ideas.
H0: To identify every possible chance of development, efficient coding skills are needed and the algorithm has to be capable of acquiring all the modifications that are needed to develop the algorithm. This can be time taking.
H2: If every possible chance can be detected to develop the following algorithm, the algorithm will be more efficient to acquire all the modifications.
H0: Literature gap will make the authors and readers understand the difficulties of the project. It will also show the gaps of the study and the research.
H3: The literature gaps may help the readers to understand the possible gaps in the research and the literature of the research.
1.6 Research Rationale
Researchers and the data scientists have got excellent results using the machine learning models and AI systems especially after using the deep learning models. This research is going to be on finding new ways to develop existing machine learning models by designing an algorithm using Python (Zhao et al. 2017). As AI systems are used in every field nowadays, whether it is academic research, industrial fields, medical technology, AI technologies have always proven to be efficient in doing particular things. Using AI technology, nothing is left to discover in the human body. Whether it is about emotions, or it is about detecting the beta signals of the human brain Ai techniques have always proverb to be the key thing to detecting the faults in the human bodies. A new algorithm will provide new datasets to determine new methodologies in the field of machine learning. Using the new datasets new machine learning techniques may evolve which can be used in the various fields.
1.7 Research Problems
There are several problems while researching this topic. The most common problem that the researchers have faced while researching on this topic is understanding the main processes which need to be automated (Rauber, et al,2020). Nowadays, it is becoming difficult to detect the facts of machine learning. The Researchers have to determine which platform should be used and what problems are to be solved. In this method also problems have arisen while doing the research on this topic.
The second problem that the authors have faced while doing the research is they did not have the sufficient data or the quality of the data was not up to that mark. To make new algorithms AI developers invest a good quality of time but if they do not have the exact data it would not be easy to make the algorithm. The same problem has been faced by the authors.
The authors did not have the adequate infrastructure to implement the ideas of machine learning (Nishino, et al,2017). The authors have used the legacy systems which often cannot handle the type of work and the workload of the algorithm.
In this chapter a brief concept about the subject is introduced. This chapter also includes the aim of the research and the objectives of this research work. This research work is based on development of machine learning models. This section has introduced the requirements of this project as well as the expectations of this research work. This section also has information regarding this subject. The questions that have been raised due to the unavailability of certain topics and the limitations of the project have been discussed. This section consists of the overview about the subject and research questions.
1.9 Research Framework
Figure: Research framework
(Source: Self created)
Chapter 2: Literature Review
Software systems are with the intelligence component that are based on the machine learning techniques that have been the widely development and now this is applied in some several numbers of fields in many field like in the commerce of electronic and in the finance systems and in some healthcares and in the entertainments machine learning is using a good way. And in the current days the practical applications that are based on the machine learning methodology that have been significant in the technique of the machine learning methodology. And the software’s that is used in the technique of the development of machine learning. The technique of machine learning has been researched and has published a topic with a broad range. In particular, all the breakthroughs in the research of deep learning and this is the force of driving and the force that is working behind the advance of machine learning. There are many papers which are written about the techniques of machine learning and the algorithms of machine learning and some applications and as well as it has been published extensively. However there are the systematic development and deployments and the operations of machine learning applications that face some major difficulties. All the methodologies and the tools of that software engineer have a great contribution to the wide range of activities of the lifecycle in some traditional information systems.
2.2 Conceptual Framework
Figure: Conceptual framework
(Source: Self created)
2.3 Empirical Study
Accordion to Kumeno and Fumihiro, 2020 methodologies and the tools of the software’s that are used in software engineering having a great contribution to the wider range to the activity if the traditional information of machine language applications because in the project of machine learning applications all the traditional systems of software’s that differ in some fundamental ways (Berg et al. 2019). A machine learning application that is involved in the computational model at least one time which has been trained in some data of training also the process of additional data in order to make the interfaces of the additional data. The behavior of the machine language model that is based on that program and that is depending on the training of data and sometimes it is unpredictable. These are the phenomenon that introduces the various kinds of uncertainties of the machine learning applications and also it differs from the software processes which are traditional.
Figure: workflow example of supervised machine learning applications
The above figure is showing the application of machine learning which is supervised. And there is a diagram of a simple workflow of the supervised machine learning application and the workflow consists of analysis of the requirements and the works that are data oriented and with the works of model orientation and the works of DevOps workflows. The requirements of performing the analysis and the activity of data analysis of the current system that requires and the works of data orientations. The oriented data works that includes all the collection of the data and in the validation of data as well as in data cleaning and also in the extraction of features (Petegrosso and Kuang, 2020). The orientations of models that actually work on cover of the model construction and design, in the training model also in the evolution and in the optimizations. And in the final stage the Devps works in all the activities to work on. And as the model of the deployment and control and monitoring and the retraining. The workflow of feedback loops that of back and the works that has been performed before the monsterization, and the training model that goes to the back loop for the feature of extraction. Machine learning is a model and the algorithm that is related to the new challenges and in such situations that make more difficult for the machine learning applications and practices for the application. Some various challenges of software engineering have been provided in the machine learning applications to cover the topic of wide ranged similar applications. Machine learning application has been challenged from SE that has been discussed in many journals and in some papers (Atalay and Çelik, 2017). But according to the knowledge of Kumeno and Fumihiro there are no papers that have been found in the paper to clarify the overview challenge for the machine learning language applications which are the SE challenge for the machine language applications that has been challenged in SE. And in which the topic of research is related in a very close way and in each of the challenges. And the knowledge body of software engineering classifies the software works and the topics into the knowledge of the comprehensive framework that is helpful in order to seek the knowledge areas. Kumeno and Fumihiro have been presumed that the framework of comprehension is helpful to seek the answers in the above research that has been discussed in the researched questions of the introduction part of the dissertation part (Drude, et al,2018). With the use of some frequent keywords that appear in each keyword and firstly they have been appearing in the knowledge area and all the machine learning languages are the keywords that have been collected in some papers and the challenge has been mapped in some topic to the knowledge of the Swebok areas. In the report of the preliminary result of the work, another part has been provided in a research method and the other section will discuss the data analysis part of the entire dissertation part. And the final chapter will contain the application of machine learning about the conclusion and recommendation with some future work and the in future the scope of this machine learning application.
According to Rasheed Ahmed and Izzat Alsmadi, the internet things are consisting of the connection of devices that actually can receive and send the data by using the simple internet connection or it may be said that all processes are working over the connection of the internet. These devices can be found in many sectors in the current days like in the hospital sectors and in the offices as well as in the telecommunication and in transportation systems. Also agriculture is taking the facility of machine learning technologies (Lehr and Ohm, 2017). All the internet used devices are growing day by day in a rapid way. A big difference is generated in the daily lives and some help is provided in the industries like the transportation system and in healthcare organizations. According to the findings of the business and with the experiment of Business insiders of 2020 and according to the report this can be said that a total market of more than 2.4 trillion Australian dollars will be the turnover in 2027. And that will include in this that there will be more than eight billion devices which will take advantage of the internet applications. Also it has been experimented that there will be more than forty two billion device users by the year 2027. And for many years the internet has brought significant benefits in society. And not only in society it also has made impact in technologies too as well as this is not too mature to provide the proper security to the devices and in communications (Zaki and Meira, 2019). The increasing factors in the connected devices that provide adversaries and also some more options in order to gain the access to all devices and to use those in order to launch the large scale of attacks. This became a challenge to the manufactures and to the consumers to provide security to the internet things or in the internet technologies are going to happen challenging in growing. Storing the data in online platforms and those can be weeks and default that may not contain any password; these are the main challenging thing to identify to the researchers. Sometimes the internet of things is shipped with some easy memorization and either default or sometimes the password may not contain with those internet things (Beam and Kohane, 2018). So that is a major problem thus the hackers can easily access the vulnerability after gaining all the access to those devices. In such vulnerability that puts the privacy of customers at risk and the hackers are allowed to use the internet-used things that are connected in those devices to the large scale attacks as similar as the DDoS. It has been reported that some medical records that are encrypted show that there are over more than five million patients in the United States or in the larger worldwide which is having the availability of the internet. And the data has been sorted online in more than one hundred and eighty seven servers and the simple can be accessed by anyone after running the simple query on that web browser. According to the report of Semantics there are some top passwords which are easily guessable. Also there are some tricks to use the internet used in devices which is shown in the below figure.
Figure: used graph
Above figure is showing how the internet used things has been improved in the society and their percentage of use in each sectors as per shown in the figure there are almost forty one percent of the internet use that advantage is taken by the health care organizations and in the manufacturing industries there are the total use of thirty three percent of internet things as well as in the all other sectors there are the use of all security and in the urban manufactures are using in four percent usages, as well as the electricity industry is taking the advantage of almost seven percent and there also some industries, those are using in a small percentage like one or two percent.
In the Symantec use of uses in the honey pots in order to analyze the behavior of adversary and in track of worldwide things of the internet. As per the internet of Symantec’s the threat report of security and the honeypots capture of semantics, on an average, there are almost five thousands attacks per month in internet technologies devices. Ninety percent of the attack has been launched against the connected camera and with the routers (Cauteruccio et al.2019). The experiment has proved that the main source for the attacking side is china, these has been seen in the experiment that china attacks the most in the internet things or technologies, by creating some applications and peoples actually use to use the applications and once the application has been installed in the device the hackers are able to check the privacy and can see the data of that particular device, as per the research there are similarity with more of the countries like Kaiten, Mirai are in the top of the list in internet technologies attacks and also there are some more country which are facing the same problem.
Figure: Machine learning method in Literature
Interconnection of security in the internet technology devices is an important aspect to protect the consumer privacy, also the infrastructures that are critical and the websites from attacks which are on a larger scale. These are the attacks that actually use the internet technology used devices as the bots of targeted flood in the website or in the networks with some huge volume of traffic which actually exceed the capacity of their bandwidth. After much research it has been studied that there are different fields in order to protect the internet technology devices from the cyber attacks.
For the purpose of the symmetric research and review that has been provided, a comprehensive analysis of some research studies and the techniques which are used by those researchers in order to protect the internet technologies from the attacks which are large-scale.
Figure: Machine learning for Flood detection
This paper has the aim of the investigation on the research of trending applications of machine learning applications in internet technology security. A systematic approach has been adopted to evaluate the recent study and in the future trends in the technology of the internet after extracting most of the scholarity and the most relevant literature that has been published in the last two years. In the current research the main objective is to drive interest in the three research areas that are internet technology and machine learning and the security information (Cauteruccio et al.2019). All the simplification and the analysis has been implemented of three domains will lead to the entire research of the researched questions that are Why Python is needed to run this software? What will be the efficiency of the new algorithm? What aspects of the new algorithm are going to be there? How is it related to real life applications? What will be the length of the new data sets? The symmetric but extensive literature reviews the approaches and that is presented in some papers that are showing specific criteria’s to use the filter of the research that are not meeting the research goals. And it is actually helping to obtain the focus and the studies which are done in recent research in internet technologies and in the technologies of machine learning that are proposed to the attacks of large scale for impacting the technologies that are used by internet technologies. And the technology of machine learning is proposed from the attacks of those devices. While many of the surveys have been studied to perform in the internet technology of the research, they have not yet performed the process systematically or it has not been discussed that the focus of machine learning is having the deep approach of learning to detect the attacks that are large scaled. In the after lessons that will be discussed briefly about the literature survey of internet technology security (Bauer, et al, 2019). The research of the primary study that is mainly focused in the internet technology devices to make more secure of the devices and to make the privacy stronger, the service of distributed denial service has been implemented in the bonnet etc. to make the comparison with the traditional intrusion to the service of detection of the technologies and in the deep learning approaches that is actually providing the promising the results to the zero detected threats. Also this has been resolved that the extensive research of the perforation in the entire field although the nature that is evolving in the nature in a fast way of the internet technology devices of types, and the traffic patterns of the network and as well as the cyber attacks that made it challenging for the detection of intrusion of system to detect the zero attack of the day. It also has been presented with a fully detailed section of the background in order to equip the entire reader with some relevant knowledge to the navigation of the further efficiency of the paperwork. And a careful analysis has been selected of the papers in the section of six extracts proposed researches of methodologies and according to the related performances of the result on the dataset of various benchmarks (Lemenkova, 2019). All the review and the results has been presented in the other sections of answer according to the authors, and the here in the research main goal is to provide more detailed knowledge and some more prevailing methodologies that are recently adopted in order to find the recent trends and the limitations and the challenges that actually helps to build the intrusion of effective detection of system in the upcoming days.
Figure: Machine learning algorithm
Default corporate predictions play the essential roles in each of the sectors in the economy, as it has been highlighted by the crisis of global financiers to increase the credit risk. This study has reviewed the corporate defaults to predict the literature that has been viewed in the engineering of finance and in the engineering of machine learning. Hyeongjun Kim and Hoon Cho and Doojin Ryu have defined the three generations of the statistical models: like the discriminant analysis, the binary models of response and the hazard models. In addition Hyeongjun Kim and Hoon Cho and Doojin Ryu have introduced the representative methodology learning to support the vector machines, and the decision trees, as well as the algorithm of neural networks. For the both machine learning and statistical model of methodology Hyeongjun Kim and Hoon Cho and Doojin Ryu have identified keys of the study that are used in the corporate default predictions. And after making the comparison of these three methods with the findings of the literature of interdisciplinary and in the review of Hyeongjun Kim and Hoon Cho and Doojin Ryu they have suggested some tasks that are newly invented in the field of machine learning for the prediction of default predictions. In the first stage the prediction of the default corporation model should be multi periodic thus the future will show its outcomes. And the second thing that needs to be shown is, the stock price and the value of the corporate sectors that is determined by the stock market and these are the important factors in order to use the default predictions. And the final point that has been implemented by Hyeongjun Kim and Hoon Cho and Doojin Ryu that the default prediction of the model is to be able to default cause in the machine learning application.
Default forecasts of the corporate sectors are used in various fields that are the total economy across. And the corporate diagnoses of the statuses based current model of predictions and in order to establish the strategy of machine learning. And the exclusive run in the business is more stable and managed in the key indicators that are actually effective in the default risk of the corporations. The strategy of investors and they can improve the portfolio after doing the examinations and the likelihood of the default corporations. In addition, macro prudential policies can be established by the government (Fol et al. 2019). And this can be improved in the related regulations of the default predictions that are used in the regular financial way. Following the same way the default predictions are helping in the design to improve the financial system. Moreover this can be said that the employing of machine learning statistical methods and the algorithms are in the default predictions and that are cutting in an advance edge in the study of financial engineering. In recent days the global crisis of finance has increased a certain credit risk that has highlighted all the importance of this current field because of the importance and the default prediction of corporations that has been studied in an extensive way in the work of Beaver. Thus far the several structures of the models have been used in order to explain the default corporations. Merton had developed the “distance-to-default” measurement of the corporate risk using the pricing model of black sholers. In recent days this has been invented and Lando and Jessen have confirmed that there is a distance of very low in the measurement of detected default of risk corporate. They have also presented the measurement of the adjustment that has been done in the asset of stochastic volatility (Roelofs et al. 2019). A structural model has been calculated by Glover for the default cost of the expectations. Brogaard has made the approach to show the reduction of stock liquidity in the corporate risk. An argument has been made in between the author and Hillegeist that the market based and measured based scholes of black merton will perform in a better way than the discrete model of hazard. The purpose came from Duffie that the stochastic model of doubly in order to make the estimation in the corporate structure that is having the default risk. In the recent days some new models of the statistical approaches have provided for more satisfactions in the outcome result than the general models.
2.4 Theories and Models
The data amount that is extracted from the process of production has been increased in an exponential way due to the sensing technology of proliferation. When it has been processed and this has been analyzed data can be brought out in the order of some knowledge and in some valuable information from the process of manufacturing, all the production system and all equipment. In the industries all equipment is maintained in a key that is important and affects the operations of time of the equipment and all its efficiency (Al-Shabandar et al. 2017). Thus all the equipment is needed to identify the problems and in order to solve the identical problems in the process of production. Machine learning is a method that has emerged as the tool of promising in the predictive maintenance of the applications in order to prevent the failures that have made up with those applications that actually depend on appropriate choice in the method of machine learning. This can be said that the main aim of the paper is to represent the systematic literature review that has been introduced about the concept of machine learning method and that method can be applied in many sectors like the hospital sector and in detection of the food that may occur and in some performance of PdM, that will show the happened explored of current field and in performance of the current machine language technique that is the art of current state. The review that has been done focuses on some significant database and that actually provides a foundation that is useful in machine learning technologies. In the test case the main result can be taken as the opportunity and the challenge as well as is supported in the new research of the workfiels of PdM.
2.5 Literature Gap
Machine learning is the tool that has attracted attention that came from the literature review and from the organization of business from the last decade, and especially there is the advanced machine learning technique to solve the problems. There are many sectors where machine learning has been used in many sectors nowadays. All the production system and all equipment. In the industries all equipment is maintained in a key that is important and affects the operations of time of the equipment and all its efficiency (Kou et al. 2019). Thus all the equipment is needed to identify the problems and in order to solve the identical problems in the process of production. In the Comtrex and in the present study of machine learning the aims all those gaps by providing the critical challenges in the literature review, that is related to the integration of artificial intelligence. Thus all the equipment is needed to identify the problems and in order to solve the identical problems in the process of production. Machine learning is a method that has emerged as the tool of promising in the predictive maintenance of the applications in order to prevent the failures that have made up with those applications that actually depend on appropriate choice in the method of machine learning.
In the above discussion of literature review all the literature points have been discussed regarding the machine learning methodologies on internet used technical platforms. Also with the same scenerionall the problems have been discussed. The software systems come with the intelligence component based on machine learning techniques that were widely used and are now being used in some areas such as e-commerce and financial systems and in some areas of healthcare and entertainment. Used in a good way. And today, practical applications based on machine learning are important to machine learning technology. And the software used in the machine learning development technique. Machine learning technology has been researched and published on a very broad topic. In particular, all the advancements in deep learning research and this is the driving force and force behind the advancement of machine learning. There are many widely published and written essays on machine learning techniques and machine learning algorithms and some applications. However, there is the systematic development, implementation and operation of machine learning applications that face some significant difficulties. These are the basic points that have been implemented in the literature review.
Chapter 3: Methodology
Methodology of delivering complex machine learning projects requires various fields of implementations. Before starting to demonstrate the methodology there is a need at first of allocation of proper team members because the methodology requires an execution of the project within a limited time, so to work together is an issue along with different logical expressions used in the projects. That's why a methodology team is needed to set at first to execute the desired outcomes according to the requirements. Apart from this the methodology development team requires an immense knowledge on the subject. Nowadays machine learning is a great approach towards the growth of the development of the respective organization. Methodology covers an entire protocol consisting of a human skill gap with proper implementation of the functions used or the tools used to execute the outcomes of the project. Researchers have understood that it is very much needed to offer a better understanding of the AI systems, especially those which are based on machine learning. It aims to establish strong foundations for the field of work. Python is the emerging platform to this era as the interface can solve a lot of issues regarding the current day challenges. In order to build a better methodology team then there should be problems, risk for execution of the project and uncertainty towards investment from the stakeholders and so on.
3.2 Research Method
Figure: Machine learning lifecycle
There are numerous methods of research on the project. Among them, the above figure has been extracted from a reference source at which various characterizations of suitable classifications along with individual functions are specified through a schematic diagram provided. Above figure is about the flowchart diagram which involves how the functions are performing with proper utilization according to the relations (Šegota, et al, 2021). The basic functions for methodology development from machine learning through Data Analysis along with software implementation. The above figure has three main entities by which entire machine learning approaches. Project requirements and KPI, methodological goals regarding investigation and developmental model and the last section consists of model work, data work, measurement, learning and analysis. The model has a condition with a parameter when to release or not with proper call in between the list of models. The developmental section plays an efficient role as the algorithm are implement in this sect with addition of other elements. Approaches like classifier which has sections such as;
The prediction or assumption result would be high at accuracy.
And it should be simple enough to gain knowledge as per implementation. Standard multi-variable statistics results, such as disintegrate analysis or 'naive Bayes' approaches towards executable, submit decision expressions in the form of sets of negative numbers and positive numbers.
weights, and attribute-value scoring for each individual
Contributions towards an Reject versus Accept preference for each choice type which is considered as assumed independent.
Lists of numbers have less meaning for the user.
Uses of a machine learning model.
The target is to develop a method for identifying risks and it should be time efficient. Tree structured classification rule is a major part to implement for the required outcomes. Information is stored as data in the database at and according to the suitable source can retrieve, update, alter, add, and can search whatever is necessary. The information is huge and to handle it properly, databases do all the storage of the data and update each day on a regular basis as the technologies are growing faster along with specific framework projects, so development plays an effective role in all aspects of an organization.
3.3 Research Resources
To do this project various machine learning techniques have been used along with the Python software. Python has been used to learn different machine learning techniques. The main reason to use Python in machine learning is that python is easy to adapt and learn as well. The syntax of python is really easy and it helps to provide the efficient codes of python.
3.4 Data Analysis Method
To gain knowledge in machine learning there are some selected steps which are as follows;
Clustering is simple enough to understand that to bind together with the similar characteristics entities and the clustering method is invariant to the output of the information as the algorithm is running at the background (Li and Kumar, 2021. ). “K-means” is the most popular method in clustering technique which also has some features to deal with the points of the data like; each cluster has to restart again in the center, chooses randomly K centers in the data and allocation of each data points according to closest of the data points with the created centers.
Mainly the classification methods deal with assuming the condition for buying the products by the customer, it simply helps to predict according to the customers perspective view (Shah, 2019). Classification methods consist of algorithms as regression using logistic expression. Algorithms are an important part in the data structure when it comes for large logical coding as per the requirements, then knowledge in skills would be an issue (Angelopoulos, 2021). After classification the information goes to the further uses of random forests, to support vector machines, use of decision trees, neurotic nets and so on.
Natural Language processing;
It is differ from the other approaches, NLP or natural language processing is a learning method used to produce text for machine learning. The document consists of NLTK stands for Natural Language Toolkit which helps to filter the unnecessary attachments.
Transferring learning means that to deliver knowledge with the subordinates. It is also necessary to adapt with the latest technological uses (Harika et al. 2020). Transfer a fraction within the trained layers and gather them with new layers and adjust according to the design pattern to execute it. And after adding new layers the neurotic net can adapt its parameters and can function as a new one.
Grouping of methods is for organizing the data collections, one such example such as Random Forest methods are a group method that mixes various Decision Trees that have been trained with diverse data collections and as a result, the quality of aRandom Forest's predicts is higher than the quality of a single Decision Trees predictions.
Reinforcement learning or augmentation learning is to instruct or to train a game or a system. Augmentation learning AL, is a method which helps for understanding from the user side view. Example like a game of Maze where the console is set in between the mouse and the maze itself is an environment.
Neurotic nets with deep learning ;
Deep learning approaches along with implementations of neurotic nets require a lot of collections of data – and a lot of processing power. The reason is that the method involves identifying many parameters within massive frameworks which is a risk to handle. It's easy to see why deep learning practitioners require extremely powerful machines with GPUs, which stands for graphical processing units.
The Principal Component Analysis or (PCA) is the most widely used amplitude reduction method, which reduces the dimension of the fundamental space by identifying new vectors that enhance the data's 're finding. When the data's sequential correlations are strong, then the principal component analysis can directly decrease the dimension of the data without having to lose too much information.
It falls under the methodological department and it helps to predict or distinguish the numerical value based on the priority data variables. The simplest method is the linear regression mathematical expression.
3.5 Research Ethics
Ethics research is the field where inquiry of specific fields such as in health science, physics, biology and also in medicine the uses of machine learning has its aspects at every possible industry in the market. Research ethics is to display all the strategic steps (Sharma et al. 2020). Function of ethics is present in such a scenario thus goes beyond communicating rules about taking decisions about what could go wrong and right, it is simple and also involves raising the moral sensibility of the machine learning researchers involved in order that they can discover greater ethical issues regarding a respectable working model (Ma and Triantafillou, 2019). The consequences of the systems they build. It is also mandatory to the developmental team to extract out the needs of the skill gap between the co workers and the employees which is an efficient method to develop a better idea and execute and can identify the field of skills gap.
Chapter 4: Data analysis_3000
Al-Shabandar, R., Hussain, A., Laws, A., Keight, R., Lunn, J. and Radi, N., 2017, May. Machine learning approaches to predict learning outcomes in Massive open online courses. In 2017 International Joint Conference on Neural Networks (IJCNN) (pp. 713-720). IEEE.
Angelopoulos, N.P., 2021. DEVELOPMENT OF TECHNIQUES FOR THE OPTIMIZATION OF THE EXECUTION OF MACHINE LEARNING APPLICATIONS IN CLOUD AND EDGE ENVIRONMENTS (Doctoral dissertation, University of Thessaly).(Angelopoulos, 2021)
Arowolo, M.O., Adebiyi, M., Adebiyi, A.A. and OKesola, J.O., 2021. Predicting RNA-Seq data using genetic algorithm and ensemble classification algorithms. Indonesian Journal of Electrical Engineering and Computer Science, 21(2), pp.1073-1081. (Arowoio et al. 2021)
Atalay, M. and Çelik, E., 2017. Büyük veri analizinde yapay zekâ ve makine ö?renmesi uygulamalari-artificial intelligence and machine learning applications in big data analysis. Mehmet Akif Ersoy Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, 9(22), pp.155-172.
Balakrishna, S., Thirumaran, M. and Solanki, V., 2020. Machine learning based improved gaussian mixture model for IoT real-time data analysis. Ingeniería Solidaria, 16(1).(Balakrishna et al. 2020)
Bauer, M. and Garland, M., 2019, November. Legate NumPy: Accelerated and distributed array computing. In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis (pp. 1-23).
Beam, A.L. and Kohane, I.S., 2018. Big data and machine learning in health care. Jama, 319(13), pp.1317-1318.
Berg, S., Kutra, D., Kroeger, T., Straehle, C.N., Kausler, B.X., Haubold, C., Schiegg, M., Ales, J., Beier, T., Rudy, M. and Eren, K., 2019. Ilastik: interactive machine learning for (bio) image analysis. Nature Methods, 16(12), pp.1226-1232.
Bernardi, L., Mavridis, T. and Estevez, P., 2019, July. 150 successful machine learning models: 6 lessons learned at booking. com. In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (pp. 1743-1751).(Bernardi et al. 2019)
Buchaca, D., Berral, J.L., Wang, C. and Youssef, A., 2020, October. Proactive Container Auto-scaling for Cloud Native Machine Learning Services. In 2020 IEEE 13th International Conference on Cloud Computing (CLOUD) (pp. 475-479). IEEE.(Buchaca et al. 2020)
Cauteruccio, F., Fortino, G., Guerrieri, A., Liotta, A., Mocanu, D.C., Perra, C., Terracina, G. and Vega, M.T., 2019. Short-long term anomaly detection in wireless sensor networks based on machine learning and multi-parameterized edit distance. Information Fusion, 52, pp.13-30.
Chen, A., Chow, A., Davidson, A., DCunha, A., Ghodsi, A., Hong, S.A., Konwinski, A., Mewald, C., Murching, S., Nykodym, T. and Ogilvie, P., 2020, June. Developments in mlflow: A system to accelerate the machine learning lifecycle. In Proceedings of the fourth international workshop on data management for end-to-end machine learning (pp. 1-4).(Chen et al. 2020)
Daghottra, A. and Jain, D., 2021. From Humans to Robots: Machine Learning for Healthcare.(Dagottra and Jain, 2021)
Devi, C.P., Vigneswari, T., Madhumitha, J.J., Sowmiya, V. and Bavithra, V., 2021. Dwell Time Management using Machine Learning in Airport. Annals of the Romanian Society for Cell Biology, pp.899-912.(Devi et al. 2021)
DHARA, A., 2018. Sentiment Analysis of Product-Based Reviews Using Machine Learning Approaches (Doctoral dissertation, West Bengal University of Technology).(Dhara, 2018)
Dong, G. and Liu, H. eds., 2018. Feature engineering for machine learning and data analytics. CRC Press. (Dong and Liu, 2018)
Drude, L., Heymann, J., Boeddeker, C. and Haeb-Umbach, R., 2018, October. NARA-WPE: A Python package for weighted prediction error dereverberation in Numpy and Tensorflow for online and offline processing. In Speech Communication; 13th ITG-Symposium (pp. 1-5). VDE.
Fol, E., de Portugal, J.C., Franchetti, G. and Tomás, R., 2019, August. Application of machine learning to beam diagnostics. In Proc. 39th Free Electron Laser Conf (FEL '19), Hamburg, Germany (pp. 311-317).
HARIKA, T., CHOUDHARY, G.Y., REVANTH, K. and REDDY, K.R., SEMI SUPERVISED MACHINE LEARNING APPROACH FOR DDOS DETECTION.(Harika et al. 2020)
Hasnat, F., Hasan, M., Khan, N.H. and Ali, A., 2018. Text classification using machine learning algorithms (Doctoral dissertation, BRAC University).(Hasnat et al.2018)
Hoque, S., 2018. Machine learning for mining imbalanced data (Doctoral dissertation). (Hoque, 2018)
Jadon, P., 2020. A new approach for sentiment analysis to measure the users opinion by using machine learning techniques.(Jadon, 2020)
Jawandhiya, P., 2018. Hardware design for machine learning. Int. J. Artif. Intell. Appl, 9(1), pp.63-84.(Jawandhiya, 2018)
Kanade, P., Singh, S., Rajoria, S., Veer, P. and Wandile, N., 2020. Machine Learning Model for Stock Market Prediction. International Journal for Research in Applied Science and Engineering Technology.(Kanade et al. 2017)
Kara, K., Eguro, K., Zhang, C. and Alonso, G., 2018. ColumnML: Column-store machine learning with on-the-fly data transformation. Proceedings of the VLDB Endowment, 12(4), pp.348-361.(Kara et al. 2018)
Khadka, N., 2019. General Machine Learning Practices Using Python.(Khadka, 2019)
Kou, G., Chao, X., Peng, Y., Alsaadi, F.E. and Herrera-Viedma, E., 2019. Machine learning methods for systemic risk analysis in financial sectors. Technological and Economic Development of Economy, 25(5), pp.716-742.
Lehr, D. and Ohm, P., 2017. Playing with the data: what legal scholars should learn about machine learning. UCDL Rev., 51, p.653.
Lemenkova, P., 2019. Processing oceanographic data by Python libraries NumPy, SciPy and Pandas. Aquatic Research, 2(2), pp.73-91.
Li, S. and Kumar, A., 2021. Towards an Optimized GROUP BY Abstraction for Large-Scale Machine Learning. Proc. VLDB Endow, 14.(Li, 2021)
Ma, Q. and Triantafillou, P., 2019, June. Dbest: Revisiting approximate query processing engines with machine learning models. In Proceedings of the 2019 International Conference on Management of Data (pp. 1553-1570).(Ma and Triantafillou, 2019)
Martin, J.S., 2021. Implementing machine learning into industrial plant energy supervision.(Martin, 2021)
Modi, J., 2019. Detecting ransomware in encrypted network traffic using machine learning (Doctoral dissertation).(Modi, 2019)
Naik, V., 2021. Machine Learning Using Serverless Computing.(Naik, 2021)
Nimishan, S. and Shriparen, S., 2018, December. An approach to improve the performance of web proxy cache replacement using machine learning techniques. In 2018 IEEE International Conference on Information and Automation for Sustainability (ICIAfS) (pp. 1-6). IEEE.(Nimishan and Shriparen)
Nishino, R.O.Y.U.D. and Loomis, S.H.C., 2017. CuPy: A NumPy-compatible library for NVIDIA GPU calculations. 31st confernce on neural information processing systems, 151.
Palomba, G., Farasin, A. and Rossi, C., 2020, May. Sentinel-1 Flood Delineation with Supervised Machine Learning. In Proceedings of the 17th ISCRAM Conference. (Palomba et al. 2020)
Parida, S. and Panchal, B., 2018. An efficient dynamic load balancing algorithm using machine learning technique in cloud environment. International journal of scientific research in science, engineering and technology, 4(2018), pp.1184-1186. (Parinda and Panchal, 2018)
Petegrosso, R., Li, Z. and Kuang, R., 2020. Machine learning and statistical methods for clustering single-cell RNA-sequencing data. Briefings in bioinformatics, 21(4), pp.1209-1223.
Prajapati, D. and Gandhi, T., CSC 219 Machine Learning (Spring–2019).(Prajapati and Gandhi, 2021)
Raicu, I., 2019. Financial Banking Dataset for Supervised Machine Learning Classification. Informatica Economica, 23(1).(Raicu, 2019)
Raj, S., 2018. Building Chatbots with Python: Using natural language processing and machine learning. Apress.(Raj, 2018)
Rauber, J., Bethge, M. and Brendel, W., 2020. EagerPy: Writing code that works natively with PyTorch, TensorFlow, JAX, and NumPy. arXiv preprint arXiv:2008.04175.
Roelofs, R., Fridovich-Keil, S., Miller, J., Shankar, V., Hardt, M., Recht, B. and Schmidt, L., 2019, December. A meta-analysis of overfitting in machine learning. In Proceedings of the 33rd International Conference on Neural Information Processing Systems (pp. 9179-9189).
Schule, M., Lang, H., Springer, M., Kemper, A., Neumann, T. and Gunnemann, S., 2021, July. In-Database Machine Learning with SQL on GPUs. In 33rd International Conference on Scientific and Statistical Database Management (pp. 25-36). (Schdule et al. 2021)
Šegota, S.B., Lorencin, I., An?eli?, N., Štifani?, D., Musulin, J., Vlahini?, S., Šušterši?, T., Blagojevi?, A. and Car, Z., 2021. Automated Pipeline for Continual Data Gathering and Retraining of the Machine Learning-Based COVID-19 Spread Models.(Šegota et al.2021)
Sethumurugan, S., Yin, J. and Sartori, J., 2021, February. Designing a Cost-Effective Cache Replacement Policy using Machine Learning. In 2021 IEEE International Symposium on High-Performance Computer Architecture (HPCA) (pp. 291-303). IEEE.(Sethumurugan et al. 2021)
Shah, N.J., 2019. The analysis of Canada's health through social media using machine learning (Doctoral dissertation).(Shah, 2019)
Sharma, V.K., Jain, R., Malik, N., Gupta, P. and Arora, R., American Sign Language Translator Using Machine Learning.(Sharma et al. 2020)
Shirahatti, A., Yadav, V., Vadde, N., Singh, A., Mahajan, P. and Sheikh, R., PREDICTION AND PREVENTIVE AWARENESS OF CHRONIC KIDNEY DISEASE USING MACHINE LEARNING ALGORITHMS. (Shirahatti et al. 2020)
Tao, Y., Coltey, E., Wang, T., Alonso, M., Shyu, M.L., Chen, S.C., Alhaffar, H., Elias, A., Bogosian, B. and Vassigh, S., 2020, August. Confidence estimation using machine learning in immersive learning environments. In 2020 IEEE Conference on Multimedia Information Processing and Retrieval (MIPR) (pp. 247-252). IEEE.(Tao et al. 2020)
Vellido, A., 2020. The importance of interpretability and visualization in machine learning for applications in medicine and health care. Neural computing and applications, 32(24), pp.18069-18083.(Vellido, 2020)
Venkatesh, R., Balasubramanian, C. and Kaliappan, M., 2019. Development of big data predictive analytics model for disease prediction using machine learning technique. Journal of medical systems, 43(8), pp.1-8.(Venkatesh et al. 2019)
Vesireddy, P., 2020. Prediction Rating of Best Cuisines in Country Capitals Using Machine Learning Algorithms (Doctoral dissertation, Dublin, National College of Ireland). (Vesireddy, 2020)
Yavits, L. and Ginosar, R., 2017. Accelerator for sparse machine learning. IEEE Computer Architecture Letters, 17(1), pp.21-24.(Yavits and Ginosa,2017)
Zaki, M.J. and Meira Jr, W., 2019. Data Mining and Machine Learning: Fundamental Concepts and Algorithms. Cambridge University Press.
Zhang, T., Pu, J., Zhang, Y., Xia, Y., Dai, H., Zhang, J., Shao, H. and Ruan, S., eduDiag: Machine learning visual diagnosis based on student behavior and performance prediction. (Zhang et al. 2020)
Zhao, R., Luk, W., Niu, X., Shi, H. and Wang, H., 2017, July. Hardware acceleration for machine learning. In 2017 IEEE Computer Society Annual Symposium on VLSI (ISVLSI) (pp. 645-650). IEEE.
Zia-ul-Mustafa, R. and Hassan, S.A., 2019, December. Machine learning-based context aware sequential initial access in 5G mmWave systems. In 2019 IEEE Globecom Workshops (GC Wkshps) (pp. 1-6). IEEE.(Zia et al. 2019)
Ziogas, A.N., Ben-Nun, T., Schneider, T. and Hoefler, T., 2021, June. NPBench: a benchmarking suite for high-performance NumPy. In Proceedings of the ACM International Conference on Supercomputing (pp. 63-74).