MATERIALS AND METHODS

In academia, the influence of the data quality dimensions of improving processes and efficiency of data quality in information systems is explained widely (Jing-Hua et al., 2009). The main objective of this study is to introduce a methodology to; measure and assess data quality dimensions by using subjective and objective measurement. The qualitative and quantitative methods are used to analyze existing dependency among dimensions and to measure data quality dimensions. Further, data mining techniques are applied to improve database quality in databases.

Qualitative and quantitative methods: The initiative approach is chosen to select commonly used data quality dimensions to improve database quality in information systems. As researchers suggest to apply more than one method to measure data quality because data quality has multi-dimensional concept (Pipino et al., 2002; Madnick et al., 2009; Lee et al., 2002) so, our methodology consists of both subjective and objective measurements to assess the quality of data.

As, the contribution of this research is reducing inconsistencies of the data in databases, we identify and analyze problems of data inconsistencies in database for information system by developing and validating a model for data quality dimensions based on applying IQ instrument (Information Quality). The IQ instrument is used as a qualitative approach to measure and validate selected dimensions with the aim of establishing the meaning of dimensions from the view of the participants who are working with different database in information systems. Next, empirical methods as a quantitative approach apply to data which gathered from the survey and interview to compute, analyze and explore dependency among dimensions which identifies the causes of data inconsistencies. After that, we measure each dimension based on attribute identification and using specific metrics in a database.

Data mining techniques: After measuring dimensions, we use and explore data mining techniques and algorithms to compute the data quality dimensions and evaluate database quality to find inconsistence data. Data mining techniques as part of this methodology has ability to detect inaccurate, inconsistent and incomplete data in order to improve the quality in database through data cleansing and appropriate methods. Generally data preprocessing decreases errors and enhance quality process. Improve data quality inconsistency in the database by proposing and assessing rules and methods is another step to visualize best and appropriate patterns and rules on the database. Implementation and simulation for validation the rules on a database is another step that assists managers to make better decision to create and improve database quality. Finally, the last step is refining and finalizing the rules and methods and put policies and enforcement to define consistence attributes of data for future and to evaluate improvement quality frequently. The flow chart of the methodology is shown in Fig. 1.

Fig. 1:	Flow chart of methodology for measuring and assessing data quality dimensions

To conclude, our methodology has both subjective and objective assessments to find and compute the causes of the low quality in the databases which is more effective and efficient than the others’ methods.

RESULTS

This methodology provides comprehensive assessment from the view of the database provider and users to evaluate databases. It can determine data quality problems and identify which dimensions has more impact on others. Thus, after gaining a database which has good quality it can be used as a benchmark in other database and can be applied in other information systems or organizations. The proposed method can make intelligent and highly sophisticated opinions and it will carry useful changes in improving processes by minimizing the cost since redundancy and overlapping activities and avoided.

Since, there has been no model and algorithm to evaluate data quality dimensions dependency and nobody applied both empirical methods and data mining techniques to improve quality, our methodology shall be more powerful and efficient to extract, detect and improve inconsistent data.

DISCUSSION

In order to improve data quality, we should know that; important finding for data provider is not equally important for data users. Ranking data quality dimensions of the current state of the competition depend on an individual information systems. Researchers have explained different definition for each dimension based on their domain. Since, there is difficult to give an exact and a common definition for each dimension, measurement has been taken and evaluated over time (Lee et al., 2002). Dimensions are major object to measure data quality and some of them are frequently used in database community such as consistency, completeness and accuracy. In order to select suitable quality dimensions that can be assessed by our method , We look for dimensions that could be objectively and subjectively measured.

Nowadays, despite the quality process and control that exist at every stage of production (Guo et al., 2010) the improvement process still faces various problems. If these problems are not identified and modified constantly it can lead to failure process in information systems. Finding quality problems and inconsistent data through different techniques and methods is a clue to reduce future failures. Inconsistent data are required to be modified to create a standard database based on their environmental usage. Also, database provider needs to identify and remove data problems to enhance the quality. Improving processes quality increase user satisfaction and leads to encourage data provider to make better decision. So, methodologies are developed to diminish failures in information systems. In brief, low data quality negatively affect on economic, financial and business process. The purpose of this research is, to present the steps of a methodology to measure and improve data quality on databases in information systems. Qualitative and quantitative methods apply to set of dimensions and database to measure the quality of the database and then empirical methods and data mining techniques apply to improve the quality process in the database environment.

Moreover, if this methodology be applied to compute other measurable quality dimensions it can assist Information System (IS) providers to have intelligent and highly sophisticated opinions on creating databases and to make accurate decisions for implementing effective and efficient information by defining and enforcing policies.

Based on the previous studies, we identified data quality as a product quality which involves the tangible item or measure. Tangible item refers to the data that can be produced, store and use later (Kahn et al., 2002). Completeness, accuracy, consistency and timeliness are examples of data quality dimensions that can be measured in database systems and are suggested by most database provider and data users (Ge and Helfert, 2008). So, we select these four dimensions that frequently used in information quality researches. These dimensions are in the category of intrinsic and contextual dimensions and can be measured by questionnaire and specific formula. We apply IQ instrument which previously applied in different field such as industry, health care and financial (Kaiser et al., 2007) as a subjective measurement. The IQ instrument is used as a qualitative approach to validate the propose dimensions in different information system with the aim of establishing the meaning of dimensions from the view of the participants. The instrument for the quality improvement process is adopted of (Shariat Panahy et al., 2013a) which is developed based on comprehensive and extensive literature and perspective of information systems users (Shariat Panahy et al., 2013b). Then, we measure value of each dimension based on a previously approved metrics (Heinrich et al., 2007; Dong et al., 2006). In the next step of the methodology, we apply the empirical methods on data which is collected from questionnaires to validate the model and to analyze existing dependencies among data quality dimensions. Applying data mining techniques on the database, assessing rules and methods to visualize appropriate patterns and rules and, implementation and simulation of validation the rules on a database are the next steps of the methodology that help managers to make better decision to create and improve database quality. Furthermore, when we refine and finalize the rules and methods, we will enforce some policies for defining consistence attributes of the data and for evaluating improvement quality regularly for the future.

In line with this research, this study sets a methodology to assess dimensions of data quality with the intention to discover relationships among dimensions. In addition to estimate the quality of data based on subjectively techniques and objective measure, our method uses empirical methods and data mining techniques to evaluate and improve data quality process via a set of enforcement policies. So, our methodology shall be more effective and efficient than other methods to measure, analyze, evaluate and improve inconsistent data towards enhancing database quality.

CONCLUSION

This At present, significant progress has been made in research on data quality. However, there is a need to develop new methods and techniques for improving data quality in new forms that can be benchmarked for different database in information systems. In this article, we have presented a methodology to measure and improve data quality in information systems. Applying the steps of this methodology is precious effort to understand the fundamental component of assessment to improve data quality dimensions. Furthermore, an empirical method helps us to investigate the relationships among dimensions and, using data mining techniques can improve efficiency and process quality in database systems.

The key contributions of this research are:

This methodology is useful in identifying data quality problems and helps database providers to make intelligent decisions and carry useful changes in the processes. It reduces the process cost since redundancy and overlapping are avoided. However, the effect and relationship among other data quality dimensions are needed to be considered and evaluated to select the most suitable improvement process in information systems.

ACKNOWLEDGMENT

Thanks to Fundamental Research Grant Scheme (FRGS), Ministry of Higher Education (MOHE), Malaysia. (FRGS Number: 03-12-10-999FR).