1 Introduction

Growth in public familiarity with information and communication technologies (ICTs) in the world, the internet in particular, has opened up opportunities for the public sector to embrace the technologies and use them to better serve citizens. The implementation of e-government systems has been attracting increased research interest, and is believed to constitute one of the most important IT implementation and organizational change challenges of the future (Warkentin et al. 2002). Electronic government is designed as a process of interaction between government and society. Carter and Belanger (2003, 2005), Pavlou (2003), and Gefen et al. (2003) states that one important factor for the success of e-government services is the acceptance and willingness of people to use e-government services.

Venkatesh et al. (2003) defines “Performance Expectancy” as the degree to which one believes in using the system will help the person to gain performance on the job. In this concept there is a combination of variables obtained from the model of previous studies of the model acceptance and use of technology. The variables are: 1. Perceived usefulness, 2. Extrinsic Motivation, 3. Job Fit, 4. Relative advantage, and the last, Outcome Expectations. In this concept there is a combination of variables obtained from the model of previous studies of the model acceptance and use of technology. The clear explanation about performance expectance could be seen in Table 1:

Table 1. Variables in performance expectancy
Full size table

Meanwhile, Davis (1989); Adams et al. (1992) defined performance expectancy as a level where a person believes that the use of a particular subject will be able to improve the work performance of the person. Chin and Todd (1995) adds the dimension of expediency TI, which makes the work easier, rewarding, increase productivity, enhance the effectiveness of, and improve job performance. It can be concluded that a person’s trust and feel by using an information technology will be very useful and can enhance performance and job performance.

Huang (1998) states that the clustering operation is required in a number of data analysis tasks, such as unsupervised classification and data summation, as well as segmentation of large homogeneous data sets into smaller homogeneous subsets that can be easily managed, separately modeled and analyzed. Meanwhile, a well-known approach for data clustering is using rough set theory (Pawlak 1982, 1991; Pawlak and Skowron 2007). For example, Mazlack et al. (2000) had developed a rough set approach in choosing partitioning attributes. One of the successful pioneering rough clustering for categorical data techniques is Minimum–Minimum Roughness (MMR) proposed by Parmar et al. (2007).

However, pure rough set theory is not well suited for analyzing noisy information systems. A knowledge discovery system must be tolerant to the occurrence of noise. For example, in the previous work on constructing student models through mining students classification-test answer sheets by Wang and Hung (2001), much noise was found in the classification tables, either the feature values or the class values, created by students. Their empirical results showed that attention should be paid to handle the noisy information in order to reach a satisfactory prediction accuracy (Wang 2005).

In this paper we present a real dataset of the users of e-government services. This data were taken from a survey aimed to identify of citizen behavior in using e-government. Descriptive statistics is used to find out the Mean (M) and Standard Deviation (SD) to identify the potential sources of study behavior. It is ran in SPSS version 22.0 and the results show that there are 5 potential sources of study performance expectancy.

The traditional main objectives of grouping awareness citizen in using e-government service are to deal with the uncertainty due to design intervention, to conduct a treatment to reduce a lack of awareness and further to improve citizen’s public service. To achieve this objective, certain clustering techniques are also being applied. Clustering a set of objects into homogeneous classes is an important Data mining operation.

The remainder of this paper is organized as follows. Section 2 describes the related work and prose method. Section 3 describes the study’s performance expectancy of e government data set. Section 4 describes experiment result. Finally, the conclusions of this work are reported in Sect. 5.

2 Proposed Method

Using MDA.

3 The Study’s Performance Expectancy of e-Government Dataset

The data set was taken from a survey in Bandung. A total population were 200 people take part in this survey. The profile of the respondents is used to provide a description of the characteristics of the sample, so it is very useful in the discussion of the results of the study investigators. The majority of respondents were women, i.e. 105 people, and the respondents were male is as much as 95 peoples. To analysis the data, for distribution of study performance expectancy scores, it follows likert-scale, i.e., 1 very not agree; 2 not agree; and, 3 neutral; 4 Agree and 5 very agree. In this survey, the study performance expectancy questionnaire has been test for reliability with alpha score yielded 0.699 and accessing content validity. Table 2 describes each attribute of performance expectancy study include the mean, standard deviation, variance and range.

Table 2. Summary of the study’s performance expectancy of e-government dataset
Full size table

3.1 Perceived Usefulness

Perceived usefulness is a leading source with M = 4.010 and SD = 0.4701. Perceived usefulness refers to is the extent to which the person believes that using a particular system would enhance his job performance (Davis 1989). Table 3 describe data distribution include frequency and percent.

Table 3. Summary of perceived usefulness data distribution
Full size table

3.2 Extrinsic Motivation

The second source is extrinsic motivation, it refers to the perception that users would and want to do an activity because it is considered a valuable role in achieving a different result from the activity itself, such as improved job performance, earnings, or promotions (Davis et al. 1992). This variable has M = 3.680 and SD = 0.680 (Table 4).

Table 4. Summary of extrinsic motivation data distribution
Full size table

3.3 Job Fit

The third source is job fit with M = 3.675 and SD = 0,6720. This variable describes how to improve individual performance base on the system capabilities. (Thompson et al. 1991). Table 5 is the result of data distribution of job fit.

Table 5. Summary of job-fit data distribution
Full size table

3.4 Relative Advantage

The fourth source is relative advantage with M = 3.685 and SD = 0.7673. it refers to the extent to which use of an innovation is considered to be better than using its predecessor (Compeau and Higgins 1995; Compeau et al. 1999). Table 6 portray the result of distribution data.

Table 6. Summary of relative advantage data distribution
Full size table

3.5 Outcome Expectations

The last source is outcome expectations, this variable refers to dealing with the consequences of behavior, based on empirical evidence, is separated into performance expectations (job-related) and personal expectations (individual goals) (Compeau and Higgins Compeau and Higgins 1995; Compeau et al. 1999). Table 7 representative of the data distribution include frequency and percent.

Table 7. Summary of outcome expectations data distribution
Full size table

4 Experiment Results

In order to apply the proposed technique, a prototype implementation system is developed using MATLAB version 7.6.0.324 (R2008a). The algorithm is executed sequentially on a processor Intel Core 2 Duo CPUs. The total main memory is 1G and the operating system is Windows XP SP3.

There are five attributes of e government performance expectancy; Perceived Usefulness (PU), Extrinsic Motivation (EM), Job-Fit (JF), Relative Advantage (RA), Outcome Expectations (OE). The MDA result is shown in Tables 8 and 9. The selected attribute is Job-fit with the value 0.075. For attribute Job-fit, we have four clusters as follows.

Table 8. MDA results of e-government performance expectancy dataset
Full size table
Table 9. MDA results of e government performance expectancy dataset
Full size table

The visualization of the clusters is captured in below figures.

5 Conclusion

In this paper, the variable precision rough set has been used as attribute selection to study performance expectancy. The technique is based on the mean of accuracy of approximation using variable precision of attributes. We elaborate the technique approach through five of variable sources among people in Bandung, i.e., perceived usefulness, extrinsic motivation, job fit, relative advantage, and the last outcome expectations. The results show that variable precision rough set can be used to groups people in each study’s performance expectancy.