Home: in the "Home" page, we introduce some basic information about the Australian company-Telstra. Further more, we have an overview of the services that we are going to provide. At the bottom, some latest happenings are showed for the customers to learn the news about Telstra.
About: in the "About" page, we have the specific allocation of the work among our five team members: two people concentrate on the front-end work, three people focus on the back-end work. At the page's bottom, we have the pictures of our team members to make the web interesting.
Prediction: in the "Prediction" page, we basically introduce the two algorithms we intended to apply to make our prediction: SVM and Random Forests. As a matter of fact, after some deep comparison, we decided to use Random Forests which would definitely offer us a better accuracy.
Product: in the "Product" page, we can proudly show the results of our work, making the prediction: searching by ID and location to get the fault severity and in the meanwhile, marking process is proceeding. We can not only get the consequence, but also it will accurately appear on map.
Events: in the "Events" page, we have a more specific introduction of our five webpages as well as the difficulties we have gone through in the whole process. The project is indeed a really valuable class for us to acquire knowledge and we have actually obtained precious experience.
For the front-end work, we focus on applying js, css, html, jason and google map api for the construction of our own interesting webpage. At the beginning, we considered for a while whether to build an App or a website, after having some careful consideration, we eventually decided to put our attention on using the webpage which is able to contain more inforamation and realize important function sometimes. We went through a lot of trials and met with numbers of trouble. We communicated with each other and got the help from the teacher as well as the teammates. Thanks to the careful instructions and the patient cooperatiion, we could finally build a pretty nice webpage.
For the back-end work, we apply SVM and Random Forests to build our own algorithms to make the prediction in order to get the recults. In the first place, we took it into consideration that which one is better: SVM and Random Forests. We compared these two approaches, after many experiments, we found that the Random Forests could provide us with a better accuracy. Therefore, we eventually used RF as the basis. As a matter of fact, people who have used RF reached the accuracy around 69% and as a matter of fact, we could finally get the accuracy around 68%. It turned out to be very satisfactory. Owting to the explicit of the work, we could do a nice job without having many big mistakes.
If you have any precious proposals, please feel free to contact us. We would really appreciate it if your suggestion helps us make progess ^=^.