agawang base

CHAPTER V

SUMMARY OF FINDINGS, CONCLUSIONS, AND RECOMMENDATIONS

In this chapter, the researchers will discuss the summary of findings, conclusions, and recommendations analyzed by the gathered data after conducting evaluation to the respondents about the study.

Summary of Findings
The researchers were able to accomplishedall the specific objectives of the study.

A virtual keyboard application with finger-based input such as tapping method implementing N-gram Algorithm and handwriting method is the first objective of the study.The study was able to develop an android virtual keyboard application that supports both tapping method and handwriting method. Tapping method is the commonly used method in entering text which preserve the traditional way of using keyboard on both standard and virtual keyboard. The study was able to utilized the N-gram Algorithm which was used to enhance the functionality of tapping method by providing word predition. Word prediction is a behavior that helps the users to enter a word without completing the tapping method. This objective was scored in a functionality criteriaof a mean4.80 as “Excellent” based on the evaluation which indicates that the developed application was capable of using the finger-based inputs that includes tapping and handwriting method and the word prediction using N-gram Algorithm. Based of the application’s reliablity, this objective score with a mean of 4.78 as “Excellent” which exhibit excellentsabout tapping and handwriting method with their entire functional requirements that proves thisobjective was developed the way it was proposed.
Part of the study is to implement Google Speech API for speech recognition and accelerometer for gesture-based input utilizing Hidden Markov Models and which is the second objective of the study. The study was able to develop a voice gestural method which implements Google Speech API. This will let the users input words or even sequence of words just by saying clearly the words wanted to enter. The study enhanced the capability of the application by providing the trending approach of entering text by the use the automatic speech recognition. The study was also able to develop gesture-based input method which is the new approach of entering text using virtual keyboards. Gesture-based method is a process of recognizing motion patterns or strokes using accelerometer sensor device and was enhanced by utilizing Hidden Markov Models. Thus, this method will let the users enter texts by performing strokes on the air. Both methods of this objective were scored with a functionality and reliabilityof mean 4.81 which equal to “Excellent” Based on the results of the evaluation, the study was capable of implementing Google Speech API for voice gestural typing and gesture-based typing using accelerometer and utilizes Hidden Markov Models.
As for the results of the third objective of the study in which is to implement genetic algorithm for smart storing of frequently used words. The study was able to develop an automatic recording of frequently used unfamiliar words or unknown words in the user’s dictionary – a database which words were stored. The study implemented Genetic Algorithm to enhance the process and provides effective functionality on this objective. With regards to the outcomes of the evaluation pertaining in this objective, the respondents marked this method 4.66 mean as “Excellent” in functionality and with a4.82 mean as “Excellent” in reliability criteria. This proves that the study was able to implement Genetic Algorithm for smart recording of unfamiliar words that users are frequently used.
The last objective of the study in which it gives an important part to attain the most compelling keyboard amongst the available keyboard applications in the market and that is to provide a keyboard skin selection and customization.The study was able to develop a selection of keyboard skins with different interface and also the study was able to develop a personalize keyboard skin to let the users customize their own keyboard interface background. This is a process of changing the keyboards interface by selecting different layouts designed by the study and by selecting personal background image stored in the gallery or even capturing image using the device camera. Refering to the result of the evaluation, this objective was scored 4.92 which is equal to “Excellent” that states the study was capable of changing keyboard skin and customization.

Conclusions To conclude the study base upon its general and specfic objectives, the researchers was able to showcase the different scopes of the study which contributed and gave important capabilities to the different users. In the succeeding text, it will discuss how the study provide different answers and solutions for every mobile user’s questions and problems about keyboard as an input device. In every smart phones, different built-in input devices like microphone, sensor, camera, and hard and soft keyboard has always been and popularly used for triggering commands, entering text, and executing behaviors. In the continuous evolution of mobile technology, combining these built-in devices into a single application will gives useful and very important roles to each and every smart phone users, normal or disabled people. That is why the proponents developed QuadKey, a virtual keyboard application for android smart phone devices with different scopes and features different functionalities which provides most of every users needs and wants. QuadKey developed and worked with the use of different built-in devices with different process but with only one objective, to help not only normal users but also those users who are strugling in using their smart phones for entering text and executes commands. Utilizing different technologies and useful processes such as APIs, frameworks, and algorithms will improve each capabilities of this application. One of the objective of the study is to include finger-based input such as tapping and handwriting method. The study developed the two finger-based methods and implemented N-gram Algorithm for the word prediction. Word prediction is one of the new concepts of a virtual keyboard application which helps users by providing an intelligent prediction of the word/s being inputted.
The study also completed the implementation of Google Speech API which was used for Automatic Speech Recognition to support the voice typing method of the study. The voice typing method is latest and trending way of executing commands and entering text. Through the help of this approach, the study integrated it with a virtual keyboard to make it more usable and sustain the needs of disabled people like blinds using voice instead of using any finger-based methods.
To provide a complete set of solutions to every disabled person, the study used the android accelerometer sensor built-in hardware for the gesture-based method and able to utilized the Hidden Markov Models for recognizing the gesture or motion patterns performed. Through this new approach of the study in virtual keyboards, even the muted users who cannot use the voice typing and board of using the finger-based typing will able to use the motion or gesture-based typing, executing commands and entering text in the air and/or motion.
Learning the words of every user like how they typed and how they used different words is also considered as one of the most important scopes when it comes to virtual keyboard applications. That is why the study implemented the Genetic Algorithm that supports the intelligent recording of user’s word which is unknown or unfamiliar. Another features and objective of the study is to provide a keyboard skin selection and customization. The study was able to developed a list of keyboard skins to provide a new interface that suitable for the user. The study was also able to developed a personalized skin background for keyboard interface customization. With this feature, the application proves that it was focused not only on what are the user’s needs but also provide different characteristics to sustain of what are the users want.

Recommendations The researchers will provide different factors to enhance the capabilities and functionalities of the study and to increase the usability of the virtual keyboard applications by the following recommendations which were based on the results of the evaluation:
1.) Swipe Typing Method – the newest and the most compelling feature of virtual keyboards for entering text in which different third partykeyboard applications that are available in the market possessed. Through this method, the keyboard will be more competent with other android virtual keyboards subsist. With this method, it will also shows how powerful is the virtual keyboard is and exhibits the complexity of the future studies.
2.) Handwriting Method Engine – thehandwriting engine used in the study results of inaccurate recognition. To enhance and improve this method, the researchers will recommend to look for the appropriate engine which can recognize both letters and words, symbols and also numbers. Since handwriting method is probably the second method of how virtual keyboard was used, the researchers will recommend to enhance the capabilities of this keyboard not just only on entering text but also on what the users needs and the other factors that can possibly enhance the virtual keyboard.
3.) Gesture-based Method Algorithm – pertaining into motion patterns, one of the delimitations or the things to be considered is the accuracy. The researchers have utilized the Hidden Markov Models in which it helps the study return the highest probability that the motion patterns performed is somehow close to the value HMM returned. The researchers will recommend to use or implement another algorithms that can return a value more accurate than the study have.
4.) Android OS Version – the study can be use in Android Gingerbread version and higher. The researchers will recommend to develop or improve the study in which different android version can be installed and used.
5.) Device Platform – the study can only be intalled and used on android platform. The researchers will recommend to develop an application that can be used even in iOS platform or any other smartphones.