Computer Science: A Literature Review on SVM Learning Technique

This literature review provides background information on SVM learning techniques used for classification models, particularly in machine learning applications. It discusses predictive modeling techniques, including the use of machine learning algorithms for prediction and description. The review covers the application of machine learning classification for risk analysis in MRI structures, comparing SVM and Naïve Bayes methods. It also examines SVM models for gene classification using microarray expression data and their potential in improving decision-making for youth protection. Additionally, the review explores SVM methods in audio classification and image remote sensing, highlighting the optimization of SVM classification parameters and the use of classifiers like SVM-GA-SV and SVM-GA-R2W2. The review concludes by referencing the implementation of a missing youth database and the role of various agencies in addressing youth runaways and violence, noting the potential of predictive analysis and GUI-based software programs in police investigations.

Running head: COMPUTER SCIENCE
COMPUTER SCIENCE
Name of the Student
Name of the University
Author Note

1COMPUTER SCIENCE
Literature Review
In this particular chapter, all the required background information concerning the main
idea of the thesis has been given1. The thesis also discusses the SVM learning technique which is
needed for producing a classification model. This particular model also highlights aboutSVM
classifier — the information focuses on the idea of machine learning techniques which has
helped the officer in solving this case. It will ultimately help in solving various kinds of issues or
cases. The background information helps in having an idea regarding graphical interface that
provides an easy way for securing both data and enabling the use of police based services2. There
are various kinds of predictive modelling techniques which are available in the present scenario.
It is considered to be free from commercially-based software. The subsequent section of the
report mainly deals with implementation of predictive modelling technique which has been
discussed in details.
Machine learning depends on the ability of various systems to learn and implement new
techniques of human interface efficiently. Machine algorithm is mainly used for the purpose of
prediction and description. Machine learning algorithm is used for explaining in a better way
with the help of linear classifier in the machine world. A proper idea has been provided
concerning the application of machine learning classification. It is mainly used for understanding
1Mercier, G., and M. Lennon (2003), Support vector machines for hyperspectral image classification with spectral-
based kernels, in Geoscience and Remote Sensing Symposium, 2003. IGARSS’03. Proceedings. 2003 IEEE
International, vol. 1, pp. 288–290, IEEE.
2. Marr, B. (2016), A short history of machine learning-every manager should read, Forbes.URL: https://www.
Forbes. com/sites/bernardmarr/2016/02/19/a-short history-of-machine-learning-every-managershould-read.

2COMPUTER SCIENCE
the various kinds of risks related to diseases which are encountered in MRI structure. Its
application helps in analysing machine learning based algorithm by making use of the MRI
approach. The MRI structure of brain highlights that there is no direct way which is needed for
finding data from clinical practice3. It has been implemented so that machine learning based
application can be easily used for the diagnosis of MRI data. In this particular application, both
kind of SVM and Naïve Bayes can be easily used for classification of MRI data. In this specific
application, both the methods of SVM and Naïve Bayes can be easily used for responding and
prediction of treatment. In some of the general cases, MRI brain application makes use of SVM,
which provides a value between the given value of 67.6 and 90.3. MRI brain application is used
for calculating the overall accuracy for predicting treatment response. This particular study
highlights specific sets of genes which are considered to be similar to the method of expression
data4. It mainly highlights various kinds of SVM model along with a different variety of metrics.
SVM model has been implemented so it can easily classify the given genes by making use of the
proper expression. The dataset which is used in this expression comes up with genes of records
that have 79 different DNA microarrays.Kernel-based functions in the SVM model which comes
up with the best prediction for unannotated along with identifying the role of yeast genes. The
ultimate focus is on the fact that SVM methods and predictive analytics can be used for
3Kim, Y.-K., and K.-S. Na (2018), Application of machine learning classification for structural brain mri in mood
disorders: a Critical review from a clinical perspective, Progress in Neuro Psychopharmacology and Biological
Psychiatry, 80, 71–80
4: Furey, T. S., N. Cristianini, N. Duffy, D. W. Bednarski, M. Schummer, and D. Haussler (2000), Support vector
machine classification and validation of cancer tissue samples using microarray expression data, Bioinformatics, 16
(10), 906–914.

3COMPUTER SCIENCE
improving the decision for protecting youth. The kernel function of the SVM model focuses on
providing the best kind of prediction which is needed for identifying the overall purpose of
genes. The result highlighted the fact that SVM models so that they can perform four other types
of methods5. In this case, two specific kinds of decision tree have been provided.
SVM is known to be a statistical method which is needed for learning algorithm in
various zones. SVM method has been applied to multiple zones of audio classification along
with retrieval of research. Depending on the study, the issue of audio classification can be built
by SVM along with a strategy for recognising binary tree6. For this particular research, the
method of audio retrieval has been given along with a new kind of metric. In the instances of
binary tree recognition, the system focus on gathering inside audio data. The system aims to
calculate DFB (Distance from the boundary) which is needed for calculatingboth audio data and
SVM.
The ultimate output of this research focuses on the fact that audio classification which
can be effectively used from SVM along with achieving a value of low rate. The total
computation of DFB is all about having a large number of vectors.
SVM classifiers come up with various kind of examination of remote sensing of images.
In this particular experiment, proper optimization of the framework is needed for analyzing and
along with an estimation of SVM classification perimeters. The ultimate goal of SVM classifier
is all about examining the overall accuracy. It highlights the fact that system needs detection of a
5Duda, R. O., and P. E. Hart (1973), Pattern classification and scene analysis, A Wiley-Interscience Publication,
New York: Wiley, 1973.
6 Bishop, C. M., et al. (1995), Neural networks for pattern recognition, Oxford university press.

4COMPUTER SCIENCE
proper kind of subset, which is required for discrimination and overcoming of the selected issues
in SVM classifier7. It is mainly done so that a generic algorithm, like GA is appropriately used.
In this experiment, there are mostly two kinds of classifiers; SVM-GA-SV and SVM-GA-R2W2,
which are needed for controlling environment. SVM-GA-SV is compared with SVM-GA-R2W2
so that probability of discriminative feature among the noisy one and leading accuracy has been
discussed in details. In the experiment, SVM-GA-SV classifier aims to provide much better
capability for detecting noise. It ultimately leads to much upper kind of classification based
accuracy in comparison to SVM-GA-R2W2 classifier. SVM-GA-SV comes up with 43 features
by making use of SVM classifier. It is mainly applied so that it can work directly with the
available hyper-dimensional space. SVM classifier aims to provide required training data into
requiring subsets and models along with two classifiers8. The overall value of SVM-GA-SV is
considered to be around $87.66. SVM-GA-R2W2 focuses on detecting 133 features by making
use of SVM classifier. It focuses on understanding the accuracy changes in comparison to
another kind of classifier. The overall result of SVM-GA-R2W2 experiment is found to be
around 91.05. In the coming chapters, the linear classifier model has been explained along with
implementing of missing youth database.
For this particular reason, a missing individual can be easily referred to as reported the
issue of missing for the interval of a few hours to months in the given policy agency. In 2005, a
various employees of Saskatchewan police policies understood the importance of MY database.
Depending on the report of MY files, there is a large number of records which needs to be taken
7Guo, G., and S. Z. Li (2003), Content-based audio classification and retrieval by support vector machines, IEEE
transactions on Neural Networks, 14 (1), 209–215
8Bazi, Y., and F. Melgani (2006), Toward an optimal SVM classification system for hyperspectral remote sensing
images, IEEE Transactions on geoscience and remote sensing, 44 (11), 3374–3385.

5COMPUTER SCIENCE
into consideration. The dataset comes up with records of 2956 records for each of the MY cases9.
It mainly highlights the fact that individual can easily report for more than one case in the
matters of issues for runaways. Data is this particular age can be used for proper kind of
distribution along with significant numbers for missing cases for which the amount varies from 9
to 18. The ultimate result of this particular report is all about analysis for youth runaways.
Various agencies have performed two kinds of action which are needed for reducing the
overall number of runaways. The primary focus is all about shifting the monthly discussion and
helping the youth in dealing with problem risk of youth. The next step is all about providing
ongoing and technical support and services. As per the Partnership Committee of Saskatchewan,
youth come up with a risk of MP in Saskatchewan10. The main task is related to various kind of
individual for understanding risk. It is needed for understanding the overall roles and
responsibilities. MP committee of Saskatchewan comes up with a health program, community
council and publication, which is necessary for creating awareness. The health education
program focuses on on-going education that aims to provide information along with risk
associated with cases of the missing youth.
As per the topic of the researcher, an idea has been provided concerning predictive
analysis which can quickly help in improving the possible decision-making protection of youth.
In 2008, the research focuses on creating classification and regression tree analysis (CART)11.
9Pfeifer, J. (2006), Missing persons in Saskatchewan: Police policy and practice, Law Foundation ofSaskatchewan
Chair in Police Studies, University of Regina Department of Forensic and Social Psychology.
10 (2015), Saskatoon Police Service 2015-2019 Business Plan, Saskatoon Police Service (SPS), available at
https://saskatoonpolice.ca/pdf/general/Business_Plan_2015-19.pdf.
11(2007), Final Report of the Provincial Partnership Committee on Missing Persons, Saskatchewan Towards
Offering Partnership Solutions To Violence Inc. (STOPS to Violence), available at

6COMPUTER SCIENCE
CPS system aims to provide the best kind of services which is needed for families and child. The
output of this study is all about neglecting and multiplying.
Recent research has highlighted one of the significant points which are needed by the
police department. It is mainly required for understanding the data provided in the police
investigation. The police department has come across the fact that they need a reasonable and
absolute number of initiatives for dealing with violence. The police department can easily make
use of software program and analytical tool for solving various kinds of issues. A software
application mainly runs with the help of GUI (Graphical User Interface). GUI has been intended
to that it can provide a certain number of visual indications towards better kind of decision
making12. It aims to focus decision making and analytical tool which is needed for overcoming a
large number of issues. GUI can be used in the police department so that it can be easily used for
collecting information with the help of graphical icons and visual based indicators. GUI can be
easily used for tracking a different kind of processes like crimes, their types and criminal history.
It can quickly provide examples of a software program which is needed by police for solving the
given task.
Proper development of technologies related to information and communication can help
in getting faster method of generating and data sharing13. There isa large number of security
http://publications.gov.sk.ca/documents/9/30559-missing-persons-final.pdf.
12Bonny, E., L. Almond, and P. Woolnough (2016), Adult missing persons: Can an investigative framework be
generated using behavioural themes?, Journal of Investigative Psychology and Offender Profiling, 13 (3),296–312.
13Russell, J., and S. Macgill (2015), Demographics, policy, and foster care rates; a predictive analytics approach,
Children and Youth Services Review, 58, 118–126.

7COMPUTER SCIENCE
organization which can provide preventive measures for software programs. It is mainly needed
for dealing with future crimes and crime based records which are required for collecting and
analysing data with the help of crime investigation. There is a large number of data mining
techniques like machine learning, intelligent agents which are needed for classifying and
predicting design. The research has been done so that an idea can be collected with respect to
criminalsand proper informationbe extracted fromhidden network structure that is built among
the employees14. It is mainly done so that criminal investigation and forensic science can be
offered by various machine learning algorithms. The algorithms are implemented in the database
at a much faster rate without data analysis knowledge recruitment by GUI15. In this thesis paper,
the overall method of MP GUI can be easily implemented which aims to provide analysis and
reporting by the help of SPS.
14Sledjeski, E. M., L. C. Dierker, R. Brigham, and E. Breslin (2008), The use of risk assessment to predict recurrent
maltreatment: A classification and regression tree analysis (cart), Prevention science, 9 (1), 28–37.
15 Brookes, D., and D. Webster (1999), Child welfare in the united states: Policy, practice and innovations in service
delivery, International Journal of Social Welfare, 8 (4), 297–307.

8COMPUTER SCIENCE
References
1. Mercier, G., and M. Lennon (2003), Support vector machines for hyperspectral image
classification with spectral-based kernels, in Geoscience and Remote Sensing Symposium, 2003.
IGARSS’03. Proceedings. 2003 IEEE International, vol. 1, pp. 288–290, IEEE.
2.Marr,B. (2016), A short history of machine learning-every manager should read, Forbes.URL:
https://www. forbes. com/sites/bernardmarr/2016/02/19/a-shorthistory-of-machine-learning-
every-managershould-read.
3. Kim, Y.-K., and K.-S. Na (2018), Application of machine learning classification for structural
brain mri in mood disorders: a Critical review from a clinical perspective, Progress in Neuro
Psychopharmacology and Biological Psychiatry, 80, 71–80.
4. Furey, T. S., N. Cristianini, N. Duffy, D. W. Bednarski, M. Schummer, and D. Haussler
(2000), Support vector machine classification and validation of cancer tissue samples using
microarray expression data, Bioinformatics, 16 (10), 906–914.
5. Duda, R. O., and P. E. Hart (1973), Pattern classification and scene analysis, A Wiley-
Interscience Publication, New York: Wiley, 1973.
6. Bishop, C. M., et al. (1995), Neural networks for pattern recognition, Oxford university press.
7.Guo, G., and S. Z. Li (2003), Content-based audio classification and retrieval by support vector
machines, IEEE transactions on Neural Networks, 14 (1), 209–215.
8. Bazi, Y., and F. Melgani (2006), Toward an optimal SVM classification system for
hyperspectral remote sensing images, IEEE Transactions on geoscience and remote sensing, 44
(11), 3374–3385.

9COMPUTER SCIENCE
9. Pfeifer, J. (2006), Missing persons in Saskatchewan: Police policy and practice, Law
Foundation of Saskatchewan Chair in Police Studies, University of Regina Department of
Forensic and Social Psychology.
10. (2015), Saskatoon Police Service 2015-2019 Business Plan, Saskatoon Police Service (SPS),
available at https://saskatoonpolice.ca/pdf/general/Business_Plan_2015-19.pdf.
11. (2007), Final Report of the Provincial Partnership Committee on Missing Persons,
Saskatchewan Towards Offering Partnership Solutions To Violence Inc. (STOPS to Violence),
available at http://publications.gov.sk.ca/documents/9/30559-missing-persons-final.pdf.
12. Bonny, E., L. Almond, and P. Woolnough (2016), Adult missing persons: Can an
investigative framework be generated using behavioural themes?, Journal of Investigative
Psychology and Offender Profiling, 13 (3),296–312.
13. Russell, J., and S. Macgill (2015), Demographics, policy, and foster care rates; a predictive
analytics approach, Children and Youth Services Review, 58, 118–126.
14. Sledjeski, E. M., L. C. Dierker, R. Brigham, and E. Breslin (2008), The use of risk
assessment to predict recurrent maltreatment: A classification and regression tree analysis (cart),
Prevention science, 9 (1), 28–37.
15. Brookes, D., and D. Webster (1999), Child welfare in the united states: Policy, practice and
innovations in service delivery, International Journal of Social Welfare, 8 (4), 297–307.