Assess data reliability from a set of criteria using the theory of belief functions

International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 6, Issue 5, May (2015), pp. 13-23© IAEME
13
ASSESS DATA RELIABILITY FROM A SET OF CRITERIA
USING THE THEORY OF BELIEF FUNCTIONS
Saloni Niranjan Shah
Department of Computer Engineering,
RMD Sinhgad School of Engineering, Savitribai Phule Pune University,
Warje, Pune-58, India
Prof. Vina M. Lomte
Assistant Professor, Department of Computer Engineering,
RMD Sinhgad School of Engineering, Savitribai Phule Pune University,
Warje, Pune-58, India
ABSTRACT
To combine information source reliability in an uncertainty representation there are many
available methods, but there are very small work focusing on the problem of evaluating the
reliability. However, in data warehousing system data reliability and confidence are very necessary
components as they are very effective for retrieval and analysis of data. Customizable criteria and
very useful down to earth decisions are provided. Even if the method is very general, it provides
more specifically interest in scientific experimental data. The method diagnosis and measure the data
reliability from a set of general criteria. It believes on the use of basic probabilistic assignments and
of evoked belief functions, since they offer a good settlement between flexibility and computational
tractability. The goal of the work is to propose a partly automatic decision-support system to help in
data reliability.
Keywords: Belief Functions, Evidence, Maximal Coherent Subsets, Trust, Data Quality, Ontology
1. INTRODUCTION
Data Reliability means that data are fairly complete and accurate to meet the intended
purposes. Data reliabilitymeans that exists when data is sufficiently complete and error free to be
crediblefor its purpose and context. Data reliability is the accuracy and completeness of computer-
processed data, given the uses they are intended for.
INTERNATIONAL JOURNAL OF COMPUTER ENGINEERING &
TECHNOLOGY (IJCET)
ISSN 0976 – 6367(Print)
ISSN 0976 – 6375(Online)
Volume 6, Issue 5, May (2015), pp. 13-23
© IAEME: www.iaeme.com/IJCET.asp
Journal Impact Factor (2015): 8.9958 (Calculated by GISI)
www.jifactor.com
IJCET
© I A E M E

14
Measuring data reliability can be leaded reviewing existing information about the data,
executing tests on the data, considering advanced electronic analysis; marking to and from source
documents; and reviewing selected system controls. To calculate data reliability is major issues for
many scientists, as these data are used in further inferences.
During collection, data reliability is mostly ensured by measurement device calibration, by
adapted experimental design and by statistical repetition.This estimation is especially important in
areas where data are scarce and difficult to obtain (e.g., for economical or technical reasons), as it is
the case, for example, in Life Sciences.
The reliability of these data depends on many different aspects and Meta information: data
source, experimental protocol. Developing generic tools to evaluate this reliability represents a true
challenge for the proper use of distributed data. In classical statistical procedures, a pre-processing
step is generally done to remove outliers.In procedures using web facilities and data warehouses, this
step is often omitted, implicit or simplistic. There are very small works present that propose a
solution to evaluate data reliability.
This paper proposes a method to assess data reliability from Meta information. Lots of
criteria are used, each one giving a piece of information about data reliability. These pieces are then
combined into a global assessment that is sent back, after proper ordering, to the end user. Such
method should handle with conflicting information, as different criteria may give conflicting
information about the reliability. It is important to be able to detect conflict and to obtain insights
about its origins, or in the absence of such conflicts, to know why such data have been declared
poorly (or highly) reliable. The method which are presenting here answers these needs, by
addressing two issues: first it provide a general approach to calculate or evaluate global reliability
from a set of criteria, second it analyze the problem of ordering the reliability assessments so that the
data are allocated in a useful manner to the end users.
The goal of the work is to propose a partly automatic decision-support system to help in a
data selection process for data reliability. As evaluating data reliability is subject to some
uncertainties, we propose to model information by the means of evidence theory, for its capacity to
model uncertainty and for its richness in fusion operators. Each criterion value is affiliated to a
reliability assessment by the means of fuzzy sets later transformed in basic belief assignments, for
the use of fuzzy sets facilitates expert removal.
This paper is organized as:
Section 2 deals with an overview of the related research regarding the data reliability
methods. Section 3 is about proposed system. Section 4 describes the features of the system. And,
Section 5 describes the conclusion.
2. RELATED WORK
In earlier days, data reliability is mostly dependent on measurement device calibration, on
adapted experimental design and on statistical repetition. In existing systems most popular system to
assess data reliability is Multi Agent System.
The approach is similar to the comparison of source assessments with reference values which
are calculatedby experts in probabilistic or possibilistic methods. It needs the definition of an
objective error function and a fair amount of data with a known reference value. This is barely
applicable in our case, as data are distributed and can be collected and stored for later use, i.e., not
having anyparticular purpose during collection of data [10].
Other approaches believe on the analysis of conflict between source information, assuming
that a source is more reliable when it presents an appropriate data. This shows that to make the
assumption that the majority opinion is more reliable. If one accepts this assumption, then the results

15
of such methods could possibly complement our approach. This assumes is not working in real life
examples as the data is dynamic [9].
In open and dynamic multi agent systems, agents often need to believe on resources or services
provided by other agents to effectuatetheir goals. During this process, agents are disclosed to the risk
of being misused by others [8].
In evidence theory, methods to calculate reliability correspondin choosing reliability scores
that minimize an error function [5].
Another paper proposes a multifaceted approach to trust models in internet environments.
The authors point out the great number of terms and twistmeanings of trust. And also point out the
difficulty to capture the wide range of subjective views of trust in single-faceted approaches. They
invoke an OWL-based ontology of trust related concepts, such as credibility, reliability, honesty,
competency or reputation, as well as a Meta model of relationships between concepts. Through
domain specific models of trust, they can invoke personalized models suited to different needs. The
idea is to provide internal trust management systems, i.e., the trust diagnosis being made inside the
system, while using the annotation power of a user community to collect trust data [4].
The methods which are presented in this paper for commendation systems is close to our
proposal, but uses possibility theory as a basis for calculation or evaluations rather than belief
functions. Another difference between this access and ours is that global information is not acquired
by a fusion of multiple uncertainty models, but by the propagation of uncertain criteria through an
aggregation function [3].
In existing system, evidence theory is used for reasoning with uncertainty, with unequivocal
connections to other frameworks such as probability, possibility and imprecise probability theories.
To deal with the conflicting information Maximal coherent subset detection algorithm is used [1].
From all above related work we can conclude that there is no security used for data collection
as well as data reliability or for documents. In our proposed system we will be included a security for
data which will be displayed to user when he will fired a query.
3. PROPOSED SYSTEM
The goal of the work is to propose a partly automatic decision-support system to help in a
data selection process for data reliability.
3.1. Design
3.1.1.Data Container
A data container is a subject-oriented, integrated, time-variant and non-volatile collection of
data in support of management's decision making process. A data container can be used to analyze a
particular subject area. For example, "sales" can be a particular subject.A data container integrates
data from multiple data sources. For example, source A and source B may have different ways of
identifying a product, but in a data container, there will be only a single way of identifying a product.
Historical data is kept in a data container. For example, one can retrieve data from 3 months, 6
months, 12 months, or even older data from a data container. This contrasts with a transactions
system, where often only the most recent data is kept. For example, a transaction system may hold
the most recent address of a customer, where a data container can hold all addresses associated with
a customer. A data container is a copy of transaction data specifically structured for query and
analysis.

16
3.1.2.Web Server
A web server is a computer system that processes requests via HTTP, the basic network
protocol used to distribute information on the World Wide Web. The term can refer either to the
entire system, or specifically to the software that accepts and supervises the HTTP requests.
3.1.2.1.Tomcat7: Apache Tomcat (or simply Tomcat, formerly also Jakarta Tomcat) is an open
source web serverand servlet container developed by the Apache Software Foundation (ASF).
Tomcat implements several Java EE specifications including Java Servlet, JavaServer Pages (JSP),
Java EL, andWeb-Socket, and provides a "pure Java" HTTP web server environment for Java code
to run in.
3.1.3. User
User access the web browser after Login into the system. He can through a query and
visualize the documents or result.
Fig.1. System Architecture
3.2. Flow Diagram
This presents the proposed system flow. As shown in the Figure 2, the flow defines the
different steps to perform data reliability.

17
Fig.2. Flow Diagram
Frist, user has to log in into the system for fetching the query. Then in next step user is
issuing the queries into the system foe finding appropriate data. Data will be search from the present
documents. After searching data, that data or document will be grouped based on the criteria present
in our paper.
Once the groups are created, apply the Maximal Coherent Subsets Detection algorithm for
removing the conflict data or document. Maximal Coherent Subsets (MCS) algorithm is a merging
strategy to deal with the problem of conflict information. When a conflicting information entering,
the MCS perform conjunctive operator on maximal subset of information, and then uses disjunctive
operation between the partial results.

18
After removing the conflicting information, Belief Function is used to assess reliable data or
document. After using the Belief Function, the will be ordered in the decreasing order to show user
most reliable data or document. The data will be shown in the form of table where first shows the
most reliable data to the list reliable data or document.
As in existing system there was not used any kind of security for the data or document. Here
we provide a security for the data and or document by applying various security algorithms.
3.3. Algorithm
Algorithm 1 Proposed Algorithm
Require: Data sets
Input: User Query
1. Enter login details
2. If(Login_validate)
3. {
4. Go to 2; //enter login details again
5. }else{
6. Check Authorization
7. Fire query
8. Create grouping of data //on the base of criteria & required query
9. Apply reliability function
10. Calculate reliability of data // base on query
11. Find Maximal Coherent Subset
12. Order data base of reliability
13. Get data //final most reliable data for query
14. }
Output: Provide Reliable Documents According to Query, Ordering by Documents reliability
3.4. Mathematical Model of a System
Let S be defined as,
S = ∑(Θ, A, L, E, I, P)
Θ: Main set of Finite ordered space.
Θ = {θ1, . . . ,θN}
A: Main set of Criteria.
A = {A1, . . . ,AS}
L: Set of linguistic terms representing the state of data reliability.
L = {very unreliable, slightly unreliable, neutral, slightly reliable, and very reliable}
E: Set of focal elements.
E = {e1, e2, e3,.......................}
I: Set of intervals.
I = {i1, i2, i3,..................}
P: Set of processes that perform the system process.
P = {P1, P2, P3, P4}

19
Step1:
P1: Set of processes for preparation of data reliability.
P1 = {a1, a2, a3, a4}
Where,
{a1=i|i is to generate the network}
{a2=j|j is to create and load the user interaction}
{a3=k|k is to create and fire query to data}
Step 2:
P2: Set of ordered space.
P2 = {ᴓ1, ᴓ2, ᴓ3, . . . . . . .,ᴓN}
Where,
N is any odd number
З ᴓi<ᴓj i< j
ᴓ1 corresponds to total unreliability
ᴓN corresponds to total reliability
Step 3:
P3: set of criteria.
P3 = {A1, A2, A3, . . . . . . . . . . ,AS}
Where,
Ai is a finite space, representing individual criteria.
We are providing the criteria group having a set of criteria and each element of Set A is treated as
individual criteria.
Step 4:
P4: set of linguistic terms and fuzzy set
P4 = {very unreliable, slightly unreliable, neutral, slightly reliable, and very reliable}
3.5. Feasibility Study
This paper is comes under NP hard because detecting maximal coherent subsets has a NP-
hard complexity.
A lot of times you can solve a problem by reducing it to a different problem. I can reduce
Problem B to Problem A if, given a solution to Problem A, I can easily construct a solution to
Problem B. In this case, easily means in polynomial time. Those problems x such that there exists an
NP- complete problem y where y Turing reduces to x.
Here the problem is algorithm defined is NP hard as we cannot predict the exact time
required for each user to get authenticated. Also how many events will occur is also hard to predict
thus the problem can be solved either by heuristics or linear programming.
3.5.1.Functional Dependencies
User is using Web Application to give their query string to obtain reliable document from data.
Input : {User Query}
Output : Provide Reliable Documents According to Query, Ordering by Documents
reliability}
Success : {Monitor reliable documents correctly according to criteria}
Failure : {Unable to find any reliable documents}

International Journal of Computer Engineering and Technology (IJCET), ISSN 0976
ISSN 0976 - 6375(Online), Volume 6, Issue 5, May (2015), pp.
3.6. Feature
1) Gives and reliable data to the user
2) And provide security for data
4. RESULTS
4.1. Input Dataset
For input dataset, we take travel agencies data in which
and User Feedback are included. In this we provide detail information of travel agencies and we are
calculating total positive and negative reviews.
4.2. Outcomes
Following figures are showing results for practical work done.
registration page is shown.
6375(Online), Volume 6, Issue 5, May (2015), pp. 13-23© IAEME
20
Gives and reliable data to the user
And provide security for data
For input dataset, we take travel agencies data in which Travel Agency
are included. In this we provide detail information of travel agencies and we are
calculating total positive and negative reviews.
Following figures are showing results for practical work done. In Fig.3 and Fig.4 Login and
Fig.3: Log in and Registration
Fig.4: Survey Expert Login
© IAEME
Travel Agency Profile, Tour Package
are included. In this we provide detail information of travel agencies and we are
In Fig.3 and Fig.4 Login and

Fig
Following fig.6 shows that when and login is done by expert or user one token will be sent to
her/his email id or mobile number and that token will be verified before he/she brows the next page.
Following Fig. shows that when a user select continent
The data is transferring from ontology server to the main application browser.
21
Fig.5: Registration of survey expert
Fig.6: Token Verification
Fig.7: Browse Tour Packages
Following Fig. shows that when a user select continent, countries will be appeared under it.
The data is transferring from ontology server to the main application browser.
© IAEME
countries will be appeared under it.

After selection of country tour packages will be displayed and will be transferred from
ontology server to application browser.
Fig. 9:
After that, user can select package and make a payment to the agency. And expert can write
their review and logout from the browser
5. CONCLUSION AND FUTURE
In Data reliability the system which is used for finding very reliable data is playing a vital
role. Existing system or methods which are used for data collection and data reliability is mostly
assure by measurement device calibration, by adapted experimen
repetition. For data collection and data reliability, we are going to use belief function and removing
conflicting information Maximal coherent subset algorith
provided in existing system here we provide the security for data sets and or documents.
Future work includes combination of current approach with other sources of information to
characterize experimental data.
22
Fig.8: selection of countries
o application browser.
Fig. 9: Tour packages for specified country
and logout from the browser.
FUTURE WORK
assure by measurement device calibration, by adapted experimental design and by statistical
For data collection and data reliability, we are going to use belief function and removing
conflicting information Maximal coherent subset algorithm is used. As studied,
system here we provide the security for data sets and or documents.
© IAEME
tal design and by statistical
For data collection and data reliability, we are going to use belief function and removing
m is used. As studied, no security will be
system here we provide the security for data sets and or documents.

23
6. ACKNOWLEDGMENT
The authors would like to express heartfelt gratitude towards the people whose help was very
useful to complete this dissertation work on the topic of “Data Reliability: Evaluate from the Theory
of Belief Functions”.
It is great privilege to express sincerest regards to P.G. Guide Prof. Vina M. Lomte as well as
H.O.D., Prof. D. N. Rewadkar, for there valuable inputs, able guidance, encouragement, whole-
hearted cooperation and constructive criticism throughout the duration of this work.
The Dissertation is based on research work in “Evaluating Data Reliability: An Evidential with
Application to Web-Enabled Data Warehouse” by SebastienDestercke were with University Paul
Sabatier, Toulouse, France, in 2008. And Patrice Buche received the PhD degree in computer
science from the University ofRennes, France, in 1990. He has been a research engineer with INRA,
Agricultural Research Institute since 2002 and an assistant professor with AgroParisTech, Paris,
since 1992. Author is very much thankful to them for such a precious work.
REFERENCES
1. SebastienDestercke, Patrice Buche and Brigitte Charnomordic, “Evaluating Data Reliability:
An Evidential with Application to Web-Enabled Data Warehouse”, IEEE transactions on
knowledge and data engineering, vol. 25, no. 1, January 2013.
2. F. Pichon, D. Dubois, and T. Denoeux, “Relevance and Truthfulness in Information Correction
and Fusion,” Int’l J. Approximate Reasoning, vol. 53, pp. 159-175, 2011.
3. Denguir-Rekik, J. Montmain, and G. Mauris, “A Possibilistic- Valued Multi- Criteria
Decision-Making Support for Marketing Activities in E-Commerce: Feedback Based
Diagnosis System”, European J. Operational Research, vol. 195, no. 3, pp. 876-888, 2009.
4. K. Quinn, D. Lewis, D. OSullivan, and V. Wade, “An Analysis of Accuracy Experiments
Carried Out over a Multi-Faceted Model of Trust”, Intl J. Information Security, vol. 8, pp. 103-
119, 2009.
5. D. Mercier, B. Quost, and T. Denoeux, “Refined Modeling of Sensor Reliability in the Belief
Function Framework Using Contextual Discounting”, Information Fusion, vol. 9, pp. 246-258,
2008.
6. Y. Gil and D. Artz, “Towards Content Trust of Web Resources, Proc. 15th Int’l Conf. World
Wide Web (WWW ’06), pp. 565-574, 2006.
7. S. Ramchurn, D. Huynh, and N. Jennings, “Trust in Multi-Agent Systems,” The Knowledge
Eng. Rev., vol. 19, pp. 1-25, 2004.
8. F. Delmotte and P. Borne, “Modeling of Reliability with Possibility Theory”, IEEE Trans.
Systems, Man, and Cybernetics A, vol. 28, no. 1, pp. 78-88, 1998.
9. R. Cooke, Experts in Uncertainty. Oxford Univ. Press, 1991. and S. Sandri, D. Dubois, and H.
Kalfsbeek, “Elicitation, Assessment and Pooling of Expert Judgments Using Possibility
Theory”, IEEE Trans. Fuzzy Systems, vol. 3, no. 3, pp. 313-335, Aug. 1995.
10. R. Cooke, Experts in Uncertainty. Oxford Univ. Press, 1991.
11. L. Zadeh, “The Concept of a Linguistic Variable and Its Application to Approximate
Reasoning-i,” Information Sciences, vol. 8, pp. 199-249, 1975.
12. Ms. Jyoti Pruthi, Dr. Ela Kumar, “Data Set Selection In Anti-Spamming Algorithm - Large or
Small” International journal of Computer Engineering & Technology (IJCET), Volume 3,
Issue 2, 2012, pp. 206 - 212, ISSN Print: 0976 – 6367, ISSN Online: 0976 – 6375.

Assess data reliability from a set of criteria using the theory of belief functions

Recommended

Recommended

More Related Content

What's hot

What's hot (19)

Similar to Assess data reliability from a set of criteria using the theory of belief functions

Similar to Assess data reliability from a set of criteria using the theory of belief functions (20)

More from IAEME Publication

More from IAEME Publication (20)

Recently uploaded

Recently uploaded (20)

Assess data reliability from a set of criteria using the theory of belief functions