CONTEXT AWARENESS FRAMEWORK FOR DEVELOPING MOBILE APPLICATIONS: A SURVEY OF SELECTED SMALL AND MEDIUM ENTERPRISES IN NAKURU TOWN, KENYA

Mobile applications act as valuable tools that can be used by entrepreneurs and small and medium enterprises (SMEs) alike to market their products/services and considerably reduce costs of doing business. The government of Kenya has invested billions of shillings in laying of fiber optic cables in Kenya, improving Kenya‘s ICT sector and coming up with projects aimed at promoting the use of technology among businesses and citizens alike. With all of these investments, there are challenges that hinder the use of business mobile applications by SMEs. Although context awareness in mobile applications is gaining recognition among mobile user and developer communities, many design scenarios still do not incorporate required context awareness as one of their primary goals. Although there are strong context awareness frameworks for mobile applications, there is still potential to improve the usability of virtual learning systems. The study established that the concept of context independence including the physical context independence and the logical context independence from the applications and mobile devices provides a graceful solution in the form of a methodological framework for the development of context-aware systems for SMEs in Nakuru Town.

collect data from the SMEs. According to Yamane (1967), the sample size was calculated using the following formula;

= 1 + ( ) 2
Where; n is the sample size, N is the population and e is the level of precision ±10 (sampling error or 90% confidence level) On applying the formula, 98.09 was the result and a sample size of 100 respondents was the number chosen to be utilized in the study. The sample size of 100 respondents from target SMEs was considered to be representative of the SMEs in Nakuru Town.

D. Instrumentation
The main tool of data collection for this study were questionnaires for primary data and secondary data from journals and journal articles, magazines, books as well as published and unpublished research works. The questionnaires were delivered by the researcher to the respondents and picked at an agreed time. After collection of the filled questionnaires and some follow-up, a total of 94 questionnaires were collected. The response rate was 94% and was considered adequate for the researcher to proceed with data analysis.

E. Data Analysis and Presentation
Data analysis refers to the process of generating value from the raw data (Johnson and Christensen, 2004).
The primary data collected were coded, edited and analyzed using Statistical Package for Social Science software (SPSS 20.0). Quantitative data collected through the questionnaire were organized and then coded.
This was done to ensure similar patterns from the answers given were identified. The collected data were checked for errors and to identify unanswered questions before the data were entered into Statistical Packages for Social Sciences (SPSS 20.0) software. Open ended questions involved qualitative data analysis at some level as the open ended questions provided responses from which themes were developed. Data analysis was carried out after data entry for both closed and open ended questions. Descriptive and inferential statistics were employed to analyze quantitative data. The statistics used include frequency percentages and the results of data analysis were presented using frequency distribution tables, pie charts and bar graphs.

A. Response Rate
A total of 100 respondents were expected to respond to the call but out of the distributed questionnaires, 94 questionnaires representing 94% were returned, but 6 questionnaires representing 6% were not returned.
The researcher did a follow up on the unreturned questionnaires and found out that out that five of the potential respondents were unavailable when the questionnaires were being collected.

B. Business Activities of the respondents
The respondents contacted in this study were asked to indicate by ticking the type of business activity that they carried out. Table 1 indicates the results obtained in relation to this question. reliable data, which comprise of the fully completed questionnaires and therefore could be in the analysis of the data. This is shown in Figure 4.1. Each option was coded with number from 1 to 5 respectively, and thereafter analyzed by calculating the mean. An extraction of the results from this evaluation is presented in Table 2.   There is thus a strong bias towards positive answers and no middle values (N). This polarization of results however leads to the data for this statement not to be statistically significant.

D. Respondents Views on Different Context Awareness Issues
The results presented here illustrate different parts of the questionnaire that addresses different aspects and themes derived from the objectives of the study. Statement one from the mobile applications usage, statements two and five concern performance, user features, speed of response, while other questions in the questionnaire handle the issues of context-awareness, user application features, context aware actors and overall impression based on user experiences with applications used.

Mobile Phone Applications Usage
The research sought to find out the mobile phone application categories that the respondents use for business. The results in Figure 3 show the kind of mobile phone applications that the users access for business purposes. The questionnaire results of this study was that mobile banking and social media was the most commonly used applications by the SMEs. 70 (82.35%) of the respondents reported mobile banking as the most common use of their mobile phone for business, and many also reported social media as their preferred method for communication via the mobile phone. The respondents also indicate that they often need people, service and location search applications in order to facilitate there day to day business activities. Of the 85 respondents in this study, 40 (47.05%) of them used people and service search applications while another 35% used location search applications. The trend is due to the ability and ease of using the internet on their phone, leading to different available applications for the users.

Frequency of Mobile Phone Applications Usage
In relation to the frequency of mobile phone applications usage, Figure 4 indicates the results from the respondents. A majority of 56 respondents representing 65.88% frequently use Mobile phone applications for business use.

Context aware Actors
The researcher also sought to determine how mobile users among the SMEs view the actions of a context aware application based on perceived test scenarios. Table 3 shows the statements obtained from the questions presented to the respondents in question five of the questionnaire, these statements helped analyze their responses as seen in Figure 5. Based on the statements indicated on Table 3 users' opinions are expressed on a five-point Likert scale and the results from the statements are indicated in Figure 5.

Context-aware Information
In terms of context-aware information the respondents were asked to take a stand in respect to first; the usefulness of contextual information for the delivery of service and secondly the user satisfaction on applications action on behalf of the user based on the contextual information, with results shown in Figure 6.
For the first statement a clear majority supported this assertion (71/85). On the second issue raised opinions are somewhat spread depending on the aspect to be achieved by the application.

E. Performance
The results in Figure 5   Appllication Used does not display aspects of its own context The application cannot be tailored The end-user cannot change internal system rules

Statements on application features
No. of respondents five support the use of contextual information to achieve better performance, flexibility, user satisfaction, speed of response and ease of access. Most questions asked have an overwhelmingly positive bias. It is however important to keep in mind that the group of user respondents were from selected SMEs in Nakuru town which implies that the results are leaned on the use of mobile applications for business. For this to be generalized, more research needs to be conducted with different groups of users.
In respect of context-awareness, some interesting features emerged in this work. The ideas from Chen et al.(2004) where they propose the use of Bluetooth to create a smart meeting room are also evident this research. They present and pursue the concepts of using context information as a basis for providing users with information and services. This research extends those ideas even further by using context not only as a factor for providing information or services, but to also use it as a basis of mobile applications acting on behalf of the user for a better experience. This feature of context aware applications is well received by the users as indicated by very positive responses to statement one.

V. Proposed Context Awareness framework
Based on the observations described in previous sections the context awareness framework architecture follows the general framework presented by (Bardram & Hansen, 2010) but makes some modification and improvements. The modified context-awareness framework consists of five layers: the device layer, the resource layer, the context layer, the storage layer, and the application layer; as shown in Figure 8.
The content of each layer is described as follows.
i. The device layer: This layer contains the physical equipments and devices operated and used in the context-aware systems including sensors, identifiers, mobile devices, and actuators, etc ii. The resource layer: Entire resources of the context aware computing environment including

VI. CONCLUSIONS
The main goal of this thesis was to conceive and design a context awareness framework to support mobile It was also evident that although mobile devices' performance is continuously improved and hardware evolved, they do fall short when the demands get high enough as described by the SMEs in Nakuru town on the basis of mobile applications ease of use, flexibility, and performance and user satisfaction. The literature review shows that considerable amount of work has previously been done in context-awareness, but the issue with existing research is that most approaches consider only one, or just a few sources of context-aware information and the efficiency of the architecture in which context acquisition in based. The answer to this as discussed in this research work is to apply a new approach to context acquisition, handling and management, a framework architecture that is based on the context database. The results and discussion were presented in chapter four. These results show a number of aspects that needs to be incorporated and addressed in context awareness framework architecture.
Objective (i) was stated as follows: To determine Small and Medium Scale Enterprises mobile phone user's experience with context aware mobile applications in Nakuru Town, Kenya; With regards to use of contextual information by currently available business mobile applications in Nakuru Town the research findings indicates that there is a reasonable level of dissatisfaction among the SMEs on the way in which these applications handle the contextual information. Those who use mobile applications among the SMEs strongly agreed that the mobile applications they have used do not present the contextual information nor does it make use of the context effectively. The respondents asserted the usefulness of contextual information for the delivery of service and secondly the user satisfaction on applications action on behalf of the user based on the contextual information. These observations justify the fact that there is need for context awareness by the application and it fulfills the goal of effective and better user experience.
Objective (ii) was stated as follows: To identify and investigate issues not addressed by currently available mobile applications context handling frameworks in Nakuru Town, Kenya. Context-aware integration and manipulation of the context as a separate component was noted to be a major aspect for effectiveness of context aware mobile applications. Identified research gaps highlighted the feasibility of applying multidimensional context-awareness in mobile applications, as the common approach today is to use a single or few dimensions of context-awareness. Such a multi-dimensional approach based on the context database manipulation to context-aware information greatly enhances the tailoring and adaptation possibilities for the applications. This is supported by the different views of the respondents on certain actions that can be taken by the mobile applications. These findings enable the study to make a conclusion that to enable the context communication and the facilitation of context information in applications the representation and ontology must be standard. The developer of context-aware applications for mobile devices does not necessarily want to be aware what is happening in the context recognition framework, he just wants to have a standard and static list of available contexts that applications may use.
Objective (iii) was stated as follows: To propose a context awareness framework in support of mobile application development for Small and Medium Enterprises in Nakuru Town, Kenya. Respondents confirmed the feasibility of the approach based on the gaps identified and highlighted a way to implement and aggregate sources of context-aware information, which, when combined, create an even stronger data foundation for the system to make weighted context-aware decisions on and to be exploited further by the mobile application regardless of their different heterogeneous platforms.
In this thesis a context-awareness framework architecture based on context database for mobile computing is proposed and developed. The main contribution in this work is to demonstrate the concept of context independence including the physical context independence and the logical context independence. The heterogeneous environments in mobile computing will gain a graceful solution. This work is intended as a starting point of future research on context-aware computing This thesis argues for a shift in focus from using context-aware technologies to design context-adaptive applications to using these technologies to enable context-aware actors. Based on this shift in focus, this research has put forward a methodological framework serving as a basis for a discussion of the possibilities for design of context-aware applications within the described approach.