Machine Learning Algorithms for Big Data

This article discusses the use of machine learning algorithms for big data analytics. It explores different methods such as supervised, unsupervised, semi-supervised, reinforcement learning, and deep learning. The article also highlights the importance of these algorithms in handling and analyzing massive volumes of data.

Running head: MACHINE LEARNING ALGORITHMS FOR BIG DATA
Machine Learning Algorithms for Big Data
Name of the Student
Name of the University
Author note

1MACHINE LEARNING ALGORITHMS FOR BIG DATA
Abstract
Big data can be defined as massive volume of data that is increasing exponentially. On the other
hand, it consists of extraction of useful information from the data through development of
possible relations among multiple data. It makes big data appearing bigger. The volume of data
dubbed as big data includes much information for a human data analyst. On the contrary,
machine learning as the service in data analytics can be helpful in managing big data in better
way. Considering the situation where it is essential to collect the massive amount of data that is
cumbersome as well as time-consuming procedure.
Keywords: Big data Analytics, Machine Learning, Learning Algorithms, supervised, semi-
supervised and reinforcement learning

2MACHINE LEARNING ALGORITHMS FOR BIG DATA
Table of Contents
Introduction..........................................................................................................................4
Availability of data..............................................................................................................4
Methods in Learning Methods.............................................................................................5
Supervised learning.............................................................................................................6
Unsupervised machine learning...........................................................................................6
Semi-supervised learning.....................................................................................................7
Reinforcement learning (RL)...............................................................................................8
Deep Learning (DL)............................................................................................................8
Result and Discussions........................................................................................................9
Conclusion...........................................................................................................................9
References..........................................................................................................................11

3MACHINE LEARNING ALGORITHMS FOR BIG DATA
Introduction
Big data analytics is becoming one of the booming research areas in computer science as
well as other industries across the world. It has obtained a success in vast as well as varied
application sectors. It includes social media, economy, health care, finance and agriculture.
Multiple intelligent machine learning techniques are designed as well as used in order to provide
big data analytics solutions. Machine learning with big data is different in several ways. While
developing successful applications of machine learning, it cannot be solely on cramming the
process of over increasing amounts of big data at algorithm as well as expecting the best possible
solutions. In the present study, availability of data, methods in Learning Methods, supervised
learning, unsupervised machine learning, semi-supervised learning, reinforcement learning (RL),
Deep Learning (DL) and results are discussed.
Availability of data
Big data analytics is considered as one of the emerging technologies as it promises to
provide better insights from big and heterogeneous data (Abadi et al. 2016). In addition, big data
analytics engages in the process of selection of the suitable big data storage as well as
computational framework that is augmented through scalable machine learning algorithms. It
involved the process of developing the technologies like sensors, electronic devices and radio
frequency IDs. In addition, cloud computing, internet of things and artificial intelligence can be
helpful to make the process easier. The technologies that are used as proper procedure for a
business issue, In addition, data streams that are produce need to be managed efficiently without
data loss. Moreover, the data is produced continuously on internet. It is considered as apparent in
the social network forms as well as discussion groups and audio and video streaming. The data

4MACHINE LEARNING ALGORITHMS FOR BIG DATA
that is available to the particular standard (Obermeyer and Emanuel 2016). On the contrary, the
data available to the standard is important for the organization. The instances of big data consist
of a retailer that can monitor product line performance online as well as predict crimes with the
use of electronic devices and sensors. These are helpful in the field of agriculture, healthcare and
manufacturing.
The organizations can collect and make query of the available data and make query of all
available data for the process of manufacturing. It is essential to consider the valuable
information that can be obtained during the process of handling multiple data streams. Big data
has brought a changed approach through allowing the managers and any user in the enterprise in
order to access huge volume of data (Chen et al. 2017). It can be possible through availability of
the tools as well as technologies like business intelligence software. Hence, development of the
process through machine learning algorithms that can be developed for modeling as well as
grouping data from the several streams as well as processing in real time in order to monitor
performance and obtain hidden insights.
Methods in Learning Methods
There are multiple ways that a person can learn machine-based algorithms. There are
learning methods that are considered as better. On the other hand, method learning belongs to
the application area of artificial intelligence where the systems are provided with specific ability
for learning as well as enhancing the process automatically from the past experience without
making programming explicitly (Meng et al. 2016). It is important to develop computer
programs in order to access as well as learn data for adjusting the actions accordingly without
having any type of human assistance. On the other hand, machine-learning algorithm can be

5MACHINE LEARNING ALGORITHMS FOR BIG DATA
divided into two popular categories such as supervised and unsupervised machine learning
algorithm.
Supervised learning
Supervised learning is considered as one of the major learning methods that are based on
learning a specific function, which maps an input to an output method based on the output based
on the examples from the process. It is important to develop the procedure with the data pair that
includes training methods. A supervised learning can be helpful in the supervised procedure and
desired output (Landset et al. 2015). It can be helpful for using the mapping the procedure. It is
used for optimal scenario and allows the procedure for mapping the new instances. An optimal
process can allow the algorithm to analyze the data generated in training as well as generate new
instances. On the other hand, an optimal scenario will allow the process for unseen examples. It
needs the learning algorithm in order to generalize from the process to learning algorithms. The
objective of supervised learning method is making proper mapping function so that the input data
as well as output data can be predicted.
It is essential to note that the supervised learning method can be widely used in machine
leaning method. There are commonly available supervised learning methods such as Support
Vector Machines (SVM), maximum entropy method (MaxENT), naïve bayes algorithm,
boosting algorithms and linear regression to handle regression problems as well as random forest
(RF) algorithm for both classification and regression (Hong et al. 2016). The issues can be
supervised learning that can fall into the groups such as classification and regression. It refers to
the issues and refer to the issues where it is important to focus on the variable.

6MACHINE LEARNING ALGORITHMS FOR BIG DATA
Unsupervised machine learning
On the other hand, in supervised learning method, the algorithm will accept the
unlabelled data to classify the process through determining the comparison between the
procedure and make it helpful to the process (LeCun, Bengio and Hinton 2015). On the contrary,
the algorithms in the method can make the procedure underlying with different types of
procedure. The algorithms involved in the method will be helpful to devise as well as discover
the hidden structure. On the contrary, it is important to select the right way that would be helpful
to discover the structure. Unsupervised learning method will be helpful from the inferences and
datasets that can explain the unlabelled data from the provided input dataset.
On the other hand, adaptive resonance theory and self-organizing maps can be helpful to
make the process better. It is also important to make the adaptive process secured for self-
organizing maps. Shanahan and Dai (2015) stated that it is important to develop the algorithm
using the association rule. The issues of unsupervised learning can be grouped as clustering as
well as association issues. The clustering process is referred as discovering the inherent groups in
the data. There is a group of integers that ends with number 7 and grouping the customers having
purchased the books (Suthaharan 2014). The organization has multiple rules and clustering the
process. The organizational rule implies the discovery of the rules that can describe large
portions of the data. For instance, people purchasing the product can be helpful in the large
portions of input data.
Semi-supervised learning
Semi-supervised learning process consists of supervised learning as well as unsupervised
learning as they can use the labeled as well as unlabeled data. In the method, training is

7MACHINE LEARNING ALGORITHMS FOR BIG DATA
important for essential as well as small-labeled data for the large amount of the unlabelled data.
It can be understood for the large input amount and the amount can be labeled through the
process (Xing et al. 2015). The group of photographs includes the images that are labeled like
tree, person as well as cat. On the other hand, labeling data is a matter of consuming as well as
extensive whereas the unlabeled data becomes east for collecting as well as storing the data.
Hence, the supervised learning process is considered as one of the important methods for
exploring as well as learning the structure in the input data. The use of supervised learning
technique can be helpful to make the best prediction for the process of supervised learning
algorithm in order to predict as well as model new input. In addition, semi-supervised learning
consists of self-training algorithms, generative models, graph-based algorithms and multi-view
algorithms.
Reinforcement learning (RL)
Reinforcement learning method can be used for discovering a specific error or give a
reward in the data during interaction of the environment. The major characteristics of
reinforcement learning can be helpful to make the action through determining an ideal behavior
in the context through using the software agents and indicate where the output is proper. The
learning method will maximize performance of a specific action. The reward signal is important
through receiving the algorithm process (Qiu et al. 2016). Q-learning algorithm, state action
reward state action and deep deterministic policy gradient are considered as policy algorithms.

8MACHINE LEARNING ALGORITHMS FOR BIG DATA
Deep Learning (DL)
Deep learning is considered as the subsets of machine learning as well as involves the
process using artificial neural networks. It is inspired with the help of functions of human brain.
It uses the machine learning techniques in order to find the solutions to solve the real world
related issues. In addition, it is effective with the help of information that demonstrates the
visualization, extraction for making automated decision and aggregation.
Result and Discussions
Machine learning methods as well as related algorithms have an important role in order to
perform calculation in the big data through a set of rules that can be applicable to every
algorithm process. Selection of appropriate algorithms overwhelm as there are several supervised
and semi-supervised algorithms procedures are involved. There is any single method that fir the
approaches used for managing the issue (Qiu et al. 2016). On the contrary, there are several
issues involved with the procedures of different algorithms. Naïve Bayes algorithm is considered
as simple and powerful algorithm method to make predictive algorithm.
Bayes theorm can be used for predicting new data using thee calculated probability
model. There are potential applications of the algorithm that are useful in the fields of genetics,
psychology and sociology. Prediction of the algorithms is developed for the data input. Hence, it
would be helpful for developing manufacturing based algorithms for optimization and predictive
maintenance. Hence, it is important to focus on the use of bayes theorm so that potential benefits
of using the analytics can be obtained.

9MACHINE LEARNING ALGORITHMS FOR BIG DATA
Conclusion
The multiple learning methods as well as related algorithms are discussed in the paper. At
current days, big data is used by most of the organizations for applicability in big data. However,
using big data in order to realize organizational advantages require the techniques, tools as well
as methods in order to harness the value from big data. There are popular datasets that could be
helpful in explaining the unlabeled data from the given input dataset. In addition, it is important
to develop the procedure that can assist to develop and cluster the algorithms that have
hierarchical clustering algorithm and density-based clustering process and k-means clustering.

10MACHINE LEARNING ALGORITHMS FOR BIG DATA
References
Abadi, M., Barham, P., Chen, J., Chen, Z., Davis, A., Dean, J., Devin, M., Ghemawat, S., Irving,
G., Isard, M. and Kudlur, M., 2016, November. Tensorflow: a system for large-scale machine
learning. In OSDI (Vol. 16, pp. 265-283).
Chen, M., Hao, Y., Hwang, K., Wang, L. and Wang, L., 2017. Disease prediction by machine
learning over big data from healthcare communities. IEEE Access, 5, pp.8869-8879.
Hong, M., Razaviyayn, M., Luo, Z.Q. and Pang, J.S., 2016. A unified algorithmic framework for
block-structured optimization involving big data: With applications in machine learning and
signal processing. IEEE Signal Processing Magazine, 33(1), pp.57-77.
Landset, S., Khoshgoftaar, T.M., Richter, A.N. and Hasanin, T., 2015. A survey of open source
tools for machine learning with big data in the Hadoop ecosystem. Journal of Big Data, 2(1),
p.24.
LeCun, Y., Bengio, Y. and Hinton, G., 2015. Deep learning. nature, 521(7553), p.436.
Meng, X., Bradley, J., Yavuz, B., Sparks, E., Venkataraman, S., Liu, D., Freeman, J., Tsai, D.B.,
Amde, M., Owen, S. and Xin, D., 2016. Mllib: Machine learning in apache spark. The Journal of
Machine Learning Research, 17(1), pp.1235-1241.
Obermeyer, Z. and Emanuel, E.J., 2016. Predicting the future—big data, machine learning, and
clinical medicine. The New England journal of medicine, 375(13), p.1216.
Qiu, J., Wu, Q., Ding, G., Xu, Y. and Feng, S., 2016. A survey of machine learning for big data
processing. EURASIP Journal on Advances in Signal Processing, 2016(1), p.67.

11MACHINE LEARNING ALGORITHMS FOR BIG DATA
Shanahan, J.G. and Dai, L., 2015, August. Large scale distributed data science using apache
spark. In Proceedings of the 21th ACM SIGKDD international conference on knowledge
discovery and data mining (pp. 2323-2324). ACM.
Suthaharan, S., 2014. Big data classification: Problems and challenges in network intrusion
prediction with machine learning. ACM SIGMETRICS Performance Evaluation Review, 41(4),
pp.70-73.
Xing, E.P., Ho, Q., Dai, W., Kim, J.K., Wei, J., Lee, S., Zheng, X., Xie, P., Kumar, A. and Yu,
Y., 2015. Petuum: A new platform for distributed machine learning on big data. IEEE
Transactions on Big Data, 1(2), pp.49-67.