Distributed feature selection for efficient

CONTACT: PRAVEEN KUMAR. L (,+91 – 9791938249)
MAIL ID: sunsid1989@gmail.com, praveen@nexgenproject.com
Web: www.nexgenproject.com, www.finalyear-ieeeprojects.com
Distributed Feature Selection for Efficient Economic Big Data Analysis
Abstract:
With the rapidly increasing popularity of economic activities, a large
amount of economic data is being collected. Althoughsuch data offers super
opportunities for economic analysis, its low-quality, high-dimensionality and
huge-volume pose great challengeson efficient analysis of economic big data. The
existing methods have primarily analyzed economic data from the perspective
ofeconometrics, which involves limited indicators and demands prior knowledge
of economists. When embracing large varieties ofeconomic factors, these
methods tend to yield unsatisfactory performance. To address the challenges, this
paper presents a newframework for efficient analysis of high-dimensional
economic big data based on innovative distributed feature selection.
Specifically,the framework combines the methods of economic feature selection
and econometric model construction to reveal the hidden patternsfor economic
development. The functionality rests on three pillars: (i) novel data pre-processing
techniques to prepare high-qualityeconomic data, (ii) an innovative distributed
feature identification solution to locate important and representative economic
indicatorsfrom multidimensional data sets, and (iii) new econometric models to
capture the hidden patterns for economic development. Theexperimental results
on the economic data collected in Dalian, China, demonstrate that our proposed

framework and methods havesuperior performance in analyzing enormous
economic data.
EXISTING SYSTEM:
However, most existing methods identifythe response factors related to
economic developmentbased on past experience and directly embody them
intoproduction function to build the correlations with economicgrowth,
overlooking the indirect effects caused by otherfactors related to them. Besides,
the existing methods relytoo much on the knowledge of economists and
embracelimited indicators and records for analysis, without fullyconsidering the
intrinsic characteristics of high-dimensionaleconomic data. Therefore, they
cannot effectively reveal theimpacts of response indictors on economic
development.
PROPOSED SYSTEM:
we explore the hidden relationsbetween economy and its response indicators
froma new angle and extract the meaningful knowledgefromeconomic big data in
order to derive right insights and conclusionsbased on an innovative distributed
feature selectionframework that integrates advanced feature selection
techniquesand econometric methods. First, in order to reducethe noise yet

promote the data quality, we propose to useusability preprocessing, relative
annual price computation,growth rate computation and normalization techniques
toclean and transform the collected economic big data. Then,to distill the
features related to economic developmentfrom high-dimensional economic data,
distributed featureselection methods are proposed to quickly partition
theimportance of given economic indicators. After that, therelations between
response indicators and economic growthcan be established by conducting
correlative and collaborativeanalysis. Our main contributions are summarized
asfollows:We present a novel framework combining distributedfeature selection
methods and econometric modelsfor efficient economic analysis, which can
revealthe valuable insights from the low-quality, highdimensionality,and huge-
volume economic big data.We develop a subtractive clustering based feature
selectionalgorithm and an attribute coordination basedclustering algorithm to
select and identify the importantfeatures of data in horizontally and
vertically.Also, weextend these two methods to distributedplatformfor economic
big data analysis.
CONCLUSION:
In this paper, we have proposed a novel feature selectionbased framework,
aiming at effective and efficientanalyzing the economic big data. In particular, it
tries tolearn the important features from the high-dimensionality,huge-volume,
and low-quality economic data for economicmodel construction. Firstly, in order

to reduce the noiseyet promote the data quality, the usability
preprocessing,relative annual price computation, growth rate computationand
normalization techniques are approached to clean andtransform the collected
economic big data. After that, a distributedsubtractiveclustering algorithm and its
improvedalgorithmare proposed to constructa two-layer featureselection model,
which selects the important features andidentifies the representative ones of
economic big data inhorizontally and vertically. With the representative
economicfactors extracted by the feature selection model, we constructthe
collaborative model for driving factors analysis ofeconomic development. Based
on the collaborative model,collaborative analysis and correlative analysis are
integratedto explore the direct and indirect relationships betweenresponse
indicators and economy. The proposed frameworkand algorithms are evaluated
on the economic developmentdata in Dalian over the past 30 years. All
experimentalresults demonstrate that our work not only accords withthe actual
development situation in Dalian, but also distillsthe hidden relations between
economy and urbanizationefficiently.
REFERENCES:
[1] A. Sheth, ”Transforming Big Data into Smart Data: Deriving Valuevia
Harnessing Volume, Variety, and Velocity Using SemanticTechniques and
Technologies,” in Proc. 30th IEEE Int. Conf. on DataEngineering, 2014, pp.2.

[2] World Economic Forum, ”Big Data, Big ImpactNew Possibilities
forInternational Development,” http : ==www3:weforum:org=docs=WEF TC MFS
BigDataBigImpact Briefing 2012:pdf,2012.
[3] ”Big Data across the Federal Government,” http
:==www:whitehouse:gov=sites=default=files=microsites=ostp=big data fact sheet
final 1:pdf, 2014.
[4] H. Giersch, ”Urban Agglomeration and Economic Growth,”Springer Science &
Business Media, 2012.
[5] R. B. Ekelund Jr and R. F. Hbert, ”A History of Economic Theoryand Method,”
Waveland Press, 2013.
[6] B. Liddle, ”The Energy, Economic Growth, Urbanization Nexusacross
Development: Evidence from Heterogeneous Panel EstimatesRobust to Cross-
sectional Dependence,” The Energy Journal,vol.34, no.2, pp.223-244, 2013.
[7] S. Ghosh and K. Kanjilal, ”Long-term Equilibrium Relationship
betweenUrbanization, Energy Consumption and Economic Activity:Empirical
Evidence from India,” Energy, vol.66, no.3, pp.24-331,2014.
[8] S. H. Law and N. Singh, ”Does Too Much Finance Harm EconomicGrowth?,”
Journal of Banking & Finance, vol.41, no.4, pp.36-44, 2014

[9] D. Baglan and E. Yoldas, ”Non-linearity in the Inflation-growthRelationship in
Developing Economies: Evidence from a SemiparametricPanel Model,” Economics
Letters, vol.125, no.1, pp.93-96, 2014.
[10] Q. Ashraf and O. Galor, ”The ’Out of Africa’ Hypothesis, HumanGenetic
Diversity, and Comparative Economic Development,” TheAmerican Economic
Review, vol.103, no.1, pp.1-46, 2013.

Distributed feature selection for efficient

More Related Content

Viewers also liked (18)

Similar to Distributed feature selection for efficient (20)

More from Nexgen Technology (20)

Recently uploaded (20)

Distributed feature selection for efficient