Ph.D.

Research proposal
Doctoral program in Computer Science

1
RESEARCH BACKGROUND

My goal in Postgraduate (Computer Science) studies is to prepare myself for career in Computer
Science research and teaching.
Later, I joined Indus University, where I chose Computer Science as Postgraduate major core
courses and curriculum have helped me to get sound grip of fundamentals of almost all aspects of
Computer Science, which is reflected by my position as among top ten students in class, I also
discovered intellectual satisfaction of solving difficult problems on my own. I spend some time
working with research scholar in Department of Computer Science that is led to research proposal,
because of this and my success in undergraduate courses.

My two-year Postgraduate in (Computer Science), that I am trying to learned core courses as much
of theory as possible in my Postgraduate program which was Cloud Computing, Data Science and
Research Methodology, as electives in the Two Year and got desired results in algorithms, under
guidance of Assoc. Prof. Dr. Zeeshan Ahmed in theory, where in I have analyzed the data of data
sets on “Home Energy Consumption in data science”. On behalf of theory I own the title of
publishing two research papers one in Data Science ( Analysis of Home Energy Consumption by
K-Mean) and (Using Naïve Bayes Algorithms to Students’ bachelor Academic Performances
Analysis). Currently, my research interests are in Human Collaboration within the field of Big
Data and Data Mining by using Data Processing and Data Modeling.

I look forward to joining your department as PhD scholar. I am confident enough that postgraduate
studies in Indus University will prove to be rewarding task and able to fulfill my career objectives.

2
INTRODUCTION

Software engineering is discipline of science which compacts with methodical improvement of

software. It has numerous approaches, process and tools to accomplish high excellence products.
The growth of big data and Internet of things (IoT) was quickly hastening and affected all regions
of technologies and commerce through cumulative profits for organizations and individuals. The
data created through IoT was main part on big data landscape which was characterizes into three
pieces: data sources, data analytics, and presentation of analytics outcomes. Characterization used
volume, variety, velocity model proposed by Gartner to define essentials of big data (Beyer, 2011).
Massive chance was offered through competence to analyzed and consume enormous sums of IoT
data, comprised in smart cities, smart transportation and network schemes, energy meters, and
isolated patient healthcare intensive care expedients .The volume’s data created through sensors,
expedients, social media, healthcare, temperature devices, and many other applications and digital
expedients that constantly produce large quantities of structured, unstructured data was intensely
cumulative that huge data generation outcomes in big data (Kambatla, 2014).

Figure 1. Architecture of big data

3
The procedures use in analytical apparatuses essential discovers patterns, trends, and associations
over diversity of period horizons in statistics (Oswal et al., 2013). When examining, these
apparatuses visualize findings in tables, graphs, and spatial diagrams for effectual decision making.
Consequently, big data analysis was serious experiment for various applications because of data
complexity and scalability of fundamental procedures that support such methods (Candela,
2012).Finally, conduct extensive simulation using Hadoop based analysis by considering simulated
scenario of social and technical network. From results and discussion, that proposed system
architecture is feasible for analyzing human behavior using big data (Ahmad, 2017).Data
producers, humans constantly generate huge amount of text data and behavior data. Joint modeling
of such diverse types of data ensures in-depth understanding of humans. On the other hand, as
knowledge consumers, different users interact with various systems to fulfill their idiosyncratic
intents. Traditional static, ad-hoc and passive machine human interactions are inadequate to
optimize such dynamic decision making process; and thus online adaptively learning driven by the
feedback from humans is desired.

Social structure of human behavior is also coming from relationship; objects (owned by a user,
such mobile, laptop, etc.) establish their relationship either by working at the same place. It is
similar to human’s way of life, e.g., human’s share personal experiences in public. These
relationships are established as a location-based application or situation-based application (Atzori,
2011).Figure 2Shows Users are using these technologies for a longer period of time that devices
are properly authorized by users and they are used to exchange data, e.g., contacts, messages,
online purchasing, etc. the behavior can be derived from the interaction of these devices. These
regularities in terms of variations and frequencies for human behavior are daily activities, such
reading/sending emails, messages, Facebook likes or sharing videos or other data, sports, online
shopping, and other work patterns are usually considered to be random.

4
 Figure 2. Human activities

OBJECTIVE

 To implement data Analytics of human activity through data mining

 To manage big data sources as expedients, social media, healthcare applications,

temperature devices, software applications and digital devices that constantly create big
amounts of structured and unstructured

 To evaluate big data Analytics of human collaboration methods are classification,

clustering, prediction association rule

LITERATURE REVIEW

Marjani et al. (2016) studies that big amounts of data were produced since past period as
contraction of Internet of things (IoT) expedients increases, however; data was not beneficial
without analytic influence. Various big data, IoT, and analytics results were facilitated persons to
acquire appreciated intuition into bulky data generated through IoT expedients. Conversely,
solutions were still in early period, and domain deficiencies inclusive assessment. It considered up-
to-date study efforts fixed to large IoT data analytics that was proposed new architecture. Finally,

5
exposed research experiments, for example confidentiality, big data mining, visualization, and
integration, were offered.

Landge and Gumaste (2015) investigated that Actual activity recognition with body sensor
networks was significant and challenging assignment. Towards capture human activity and
examine that data, both activity was very important in anomalous activity detection. It presents
automated method to activity tracking that recognizes mutual activity that happens in routine of
individual. It was binary classifier which was used to classified two data’s classes which could
track incidence of systematic actions to supervise useful health and to detect modifications in
routine of human life. It was define data mining’s activity and pursuing it in image processing
method’s arena and validate procedures on data composed from isolated sensors.

Driscoll (2010) describes Big Data as dataset that was greater than sure threshold, for instance
terabyte although others express Big Data that shows structures of big volume, velocity, and
variety. Johnson (2012) also studies very large data sets associated to customer manners, public
web posts, geo tagging, and sensor productivities by Big Data. Researchers well-defined Big Data
through volume, velocity, variety, veracity and value. Chow and Green (2013) represent social
movement in societies to data-driven decision making. Boyd and Crawford (2012) studies big data
was traditional, scientific, and scholarly realism that offered capability to collection of big quantity
of data investigate patterns, recognized financial, public, practical, and high level of extracted.
IBM (2011) torrent of attention in big data conception was quiet of infancy and an application was
extensive and diverse.
Goes (2014) suggested that was no particular big data definition, corporation managers was tried to
take steps around innovative wave, because data was rapidly fetching fresh capital’s form , diverse
currency, and new value’s source.

Maimon and Rokach (2010) studies that data mining exertion was dispersed to numerous collude
negotiator human. Impartial of data mining exertion was to outturn elucidations to attempted
problem which was treated well by certain metric, with esteem to those results that would attained
by distinct, non-collaborate negotiator. Results require assessment, contrast, and methods for
permutation as association needs conversation, and involve societies system. Establishing societies
in efficient way was non- resistant and usually needs unambiguous roles and procedures.

6
Sharma and Wang(2017) suggested that big data analytics was received significant interest in
diversity of field comprised commerce, economics, health care and IoT (Internet of Things) was
measured like significant platform in carrying public, method, statistics and things collectively in
organize to improve excellence of daily life. Still, challenge were how to efficiently extract
functional characteristic from huge amount of heterogeneous data created by source constrained
IoT devices information to offer immediate and reaction to users, and how to exploit data
conscious intellect in attractive performance of IoT networks. While cloud computing for attends
to problem in data analytics, that was individual benefits and boundaries. Researchers offered new
frame for synchronized processing cloud by incorporate reward from platform. It could utilize
knowledge and historical information accessible at cloud middle to lead edge computing
component towards pleasing diverse act necessities of heterogeneous IoT networks.

Charissis and Papanastasiou (2008) investigated innovative plan for self-propelled full-windshield
head-up display interfaces that goal to enhance driver’s dimensional sentience and retort eras with
lower discernibility settings. To accomplish these necessities, contrive and enforced occupied
archetype of human–machine interface. Certain accent was retained on prioritizing and efficacious
appearance of information accessible done with vehicular sensors, this would assistance, devoid of
disrupting, driver in efficaciously traversing vehicle in small visibility situations. Future interface is
built on simple optical appearance of actual substances to render a novel style of cooperative
regulation for motorway environments. Generally, work deliberates strategy contest of human–
machine system, enlarges on interface strategy idea and offerings result of manipulator efforts
which compared efficiency of planned HUD in contradiction of head-down display.

METHOD AND MATERIAL

Big data (BD) investigation of contains different segments as displayed in figure 3, analysis was
acknowledged much consideration, unwell implicit complexity in context of multi tenanted clusters
where some consumers’ programs execute simultaneously. BD was managed in context, which
could heterogeneous, noisy and not consist of truthful model. Toward track source, to handle
ambiguity and error issues that was decisive to achievement, and yet rarely mentioned in identical
sniff. It may need to figure out good questions based on data which is required smarter systems and
7
also better support for user interaction with analysis. Actually, main bottleneck in number of
people authorized to examine questions of data and analyzed.

Figure 3. Steps of Big data analysis

Acquisition Data: BD was not stand up beyond vacuum that was documented from data producing
source. It deliberated capacity to intelligence and detects about world, from heart rate of senior
civilian, and existence of contaminants in midair human respire, and to scheduled array shrivel,
which was yield up to one million terabytes of raw data per day. Also systematic research and
imitations could simply yield data’s petabytes today.

Extraction: Statistics composed was not layout prepared for analysis that was deliberate gathering
of automatic health records in clinic consist of copy out notations from a number of doctors,
structured data from devices and capacities, and image information. It could not sabbatical
information in form and quiet well analyzed which necessitate acts mining method that attractions
available statistics from causal sources and states in structured.
Integration: Torrent of statistics that was not sufficient merely to highest and toss it into
repository. Statistics from range of scientific experiments, repository with data sets, that was
improbable someone would forever be capable to discovery.
Modeling: Approaches aimed at query and mining huge statistics was vital diverse from traditional
analysis on minor samples. However, noisy BD was valued than small samples because overall
8
statistics obtain from repeated patterns and correlation frequently suppress distinct fluxes and
reveal more consistent concealed patterns and knowledge. Additional, interrelated BD forms big
heterogeneous statistics grids, with which data redundancy could be discovered to reward for
missing statistics, to authenticate reliable associations, to reveal intrinsic clusters, and to expose
concealed associations.

Interpretation: ability to analyze data of limited value if users could not comprehend study.
Decision maker, delivered with analysis’s outcome, has toward interpreted outcomes which was
involved investigative all assumptions made and reviewing study. Furthermore, it was various
potential bases of mistake; computer schemes could have bugs, models constantly have
assumptions, and outcomes could be founded on wrong data. Through reviewing how finest to
capture, stock, and query provenance, in aggregation with methods to arrest sufficient meta data,
that could generated substructure to delivered consumers with capability both to interpret analytical
outcomes achieved and to replication analysis with diverse assumptions and data sets.

Human Collaboration: In spite of incredible developments prepared in computational study,

remain several patterns that humans could identify but workstation algorithms require tough time
discovery.BD analysis scheme must support contribution from various humanoid specialists, and
shared result’s A Big Data analysis system must support input from multiple human experts, and
shared outcome’s exploration. Specialists were divided in space and time once expensive to collect
whole team together in area. Statistics scheme has to admit dispersed skilled input, and
collaboration while it is new technique of harness human inventiveness to solved difficulties
through gathering source. However, it was predictable to be those who contain other reason and
capability a few might have motive to provide fake information deliberate effort to deceive.
Although mistake was identified and accurate by crowd, human require tools to make easy. It was
required framework to utilize in study of such gathering source data with contradictory report.
When humans could seem at analysis of restaurant, which was optimistic and critical, and come up
among synopsis evaluation based on which could make a decision whether to try eating; that
require PC to be capable to do corresponding. Ambiguity problem and fault become yet marked in
precise form of crowd source, term participatory common sense, as all human being with mobiles
could operate like multi modal sensor gathering different kind of statistics immediately such that
picture, video, audio, location, time, speed, direction and acceleration that composed data was
9
possibly spatially and temporally correlated was exploited to enhanced appraise precision.

Figure 4. Human collaboration data analysis

DELIVERABLES AND PROGRAM SCHEDULE

The education backbone stand greatly of study effort by specific helped discussion group from
(university name) university in addition spouse talent affiliates related to distance learning
methods, feature learning, machine learning, data mining, pattern recognizing, hip addition, an
occupied stay towards two-way societies existing analysis of data assimilation methods then data
analysis modeling, data analysis processing its alteration to inferior leeway areas. The general
approach determination constructed the proceeding plans delineated and according to PhD.
Advisor. management, preprocessing and distribution will be progressive focused on approaches
specified through appropriate architecture.

The overall attitude is itemized under:

1st year:

o Coursework and current the architecture solutions with the help literature review.

o Collect the data from different aspects

o Appraisal allusion software consumer attendants connected through kit and booklets,

10
 documents.

o Literature review to use best approach for data modeling and clustering.

o Study data assimilation, parameterization schemes and down scaling techniques

o Distance learning and features learning techniques and related systems literature

o Optimization using distributed programming and required tools

2nd year:

o Proposing a representation and an algorithm for data

o Evaluation of features and extracting features from videos with the help of existing different
applications.

o Optimization using distributed programming

o Test the developed model

3d year:

o Improvement of system

o Defense with conferences

o Writing research papers

o Writing PhD manuscript

First year Second year Third year
Sep Dec Feb Ma Sep Dec Feb Ma Sep Dec Feb Ma
To To To y. To To To y. To To To y.
No Jan Apri To No Jan Apri To No Jan Apri To
v 201 l. Aug v 201 l. Aug v 202 l. Aug
201 8 2018 . 201 9 2019 . 201 0 2020 .
7 201 8 201 9 202
8 9 0
Research
Initiation
Course work
Literature Survey
Familiarization
with Tools,
Techniques,
Commercial
video data sets
11
 Learning
techniques and
algorithms for
semantic analysis.

Formulate
Hypothesis/Conduct
Experimental
Work
Journal Findings,
Papers Publishing
Compilation of
Thesis

REFERENCES

Ahmad.A, 2017, Socio-Cyber Network: The Potential of Cyber-Physical System to Define Human
Behaviors using Big Data Analytics. Future Generation Computer Systems.

Steve Lohr, 2012, The Age of Big Data. New York Times,
http://www.nytimes.com/2012/02/12/sunday-review/big-datas-impact-in-the-world.html

Kambatla, K., et al., 2014, Trends in big data analytics. Journal of Parallel and
DistributedComputing.74 (7):p.2561-2573.

Landge.A, Gumaste. S.V, 2015, Anomalous Human Activity Detection Using SVM and Active
Leaning Based Approach, IJIRCCE, 3(9)

Beyer, M., SolvingG. S, 2011, Big Data Challenge Involves More Than Just Managing Volumes of
Data.

Candela, L., D. Castelli, and P. Pagano, 2012, Managing big data through hybrid data
infrastructures. ERCIM News, 89: p. 37-38.

Atzori, Luigi, Iera. A, and Morabito. G, 2011,"Siot: Giving a social structure to the internet of
things." IEEE communications letters 15, no. 11: 1193-1195.

Oswal, S. and S. Koul. 2013, Big Data Analytic and Visualization On Mobile Devices.
inProceedings of National Conference on New Horizons in IT-NCNHIT.

12
Peter A. Chow. W, Sandy E. Green.J.R, 2013,” Data Mining Difference in the Age of Big Data:
Communication and the Social Shaping of Genome Technologies from 1998 to 2007.”
International Journal of Communication 7, 556–583

David L. Banks and Yasmin H. Said, 2006,” Data Mining in Electronic Commerce.” Statistical
Science, Vol. 21, No. 2, A Special Issue on Statistical Challenges and Opportunities in Electronic
Commerce Research, pp. 234-246.

Kurt. F and Rita.G, 2013, “Big Data: Implications for Financial Managers.” The Journal of
Corporate Accounting &Financ.

BarkerJ. F, Peter A. Chow. W, Charters. A and Dung. H, 2016, “Genomic Big Data and Privacy:
Challenges and Opportunities for Precision Medicine.” Computer Supported Cooperative Work
(CSCW) , Volume 25, Issue 2–3, pp 115–136

Teemu M, David.B, 2008, “Management control systems as a package—Opportunities, challenges

and research directions.” Management Accounting Research19( 4), Pp. 287-300.

Tawny.S and Brian.D. J, 2012, Entertainment in the Age of Big Data.Proceedings of IEEE Vol.
100.

Danah.B and Kate.C, 2012, CRITICAL QUESTIONS FOR BIG DATA, Information,
Communication & Society, 15:5, 662-679.

Paulo B. G, 2014, “Editor's comments: big data.” MIS Quarterly, 38(3).

Marjani. M, Nasaruddin. F, Gani. A, Karim.A, Hashem. I.A. T, Siddiqa. A, Yaqoob. I, 2016, Big
IoT Data Analytics: Architecture, Opportunities, and Open Research Challenges, IEEE

Charissis.V and Papanastasiou. S, 2008, Human–machine collaboration through vehicle head up

display Interface, Springer-Verlag London
O. Maimon, L. Rokach, 2010, Data Mining and Knowledge Discovery Handbook, 2nd ed.
Springer, DOI 10.1007/978-0-387-09823-4_54

Sharma. S. K, Wang. X, 2017, Live Data Analytics with Collaborative Edge and Cloud Processing
in Wireless IoT Networks, IEEE

13