The utilisation of neural networks as a data mining technique has grown significantly during the last two decades. However, there is not much known on how neural networks are actually developed. This survey based research project has been initiated to investigate how neural networks are developed.
This survey’s questions focus on how developers in industry build neural network applications and the system development methodologies that they follow. This survey should take no more than 15 - 20 minutes to complete.
This is an anonymous survey. No data will be collected that will allow any person or organisation to be identified. The only people who will have access to the data will be the chief investigator and the researcher. This survey is entirely voluntary. If you want to obtain a copy of the results, please contact the researcher as listed below.
Thank you for your participation in this survey and we hope to hear from you.
Keith Lai
Honours Student
Monash University
School of Information Technology
Monash DSS LAB
0433 206 816
[email protected] This survey’s structure is based on the steps of a typical system development process. It is divided into several sections:
Section A: Planning
Section B: Data Anaylsis
Section C: Development
Section D: The Development Methodology
Section E: Outcomes
When answering the following questions, think about the most recent neural network project you have worked on.
A. Planning
This section involves understanding how the project was resourced and planned.
1. Who initiated the project? (Please tick one) 2. What was the purpose of developing the neural network system? (e.g. Sales forecasting, stock market predictions, etc.) 3. How many people were in the project team? 4. Was there a formal requirements planning phase? 5. How did you decide whether a neural network based system was appropriate for the given problem? 6. In general, how would you rate the completeness of your planning phase?
B. Data Analysis
This section focuses on how data is prepared and collected before it is used for training the neural network.
7. Prior to data collection, was there data preparation required? 8. Did you have to significantly modify the data or its format? If yes, please list the processes involved. 9. How was data collected for training the neural network? 10. Did you try using the same set of data for training and testing the neural network? 10a. If yes, what were the results? 11. Where did the data come from? (Tick all that apply) 12. Were there often data quality problems that needed to be fixed before data was fed into the neural network for training? 13. What were the factor(s) that determined the amount of data to be collected? (Tick all that apply)
C. Development
Section C focuses the development phase of a neural network application and the techniques used during network development.
14. Was a neural network development environment used to build and train the network? 14a. If yes, please specify the neural network development environment used to build the neural network. 14b. If no, what was the technology used to build and train the neural network? 15. Did you use the same neural network development environment for deploying the application? 15a. If no, what other development environment was used for deployment of the application? 16. Which neural network architecture is normally used for training? (E.g. Hopfield, Kohonen, etc.) 16a. Please describe the neural network architecture. 17. Why did the project team choose the neural network architecture? (Please tick one) 18. How many iterations did it take to train the neural network? (Please tick one) 19. In your project, what were the factor(s) in determining the technology used for implementing the neural network system? (E.g. based on requirements, costs, applicability of the technology, etc.). 20. What platform was the neural network system based on after implementation? (Tick all that apply) 21. What operating system does the neural network system run on? (Please tick one)
D. The Development Methodology
This section discusses the development methodology that is used to build a neural network application.
22. Did you follow a specific development methodology in building the neural network system? 22a. If yes, please describe. 23. Was the development process similar to conventional methodologies such as the SDLC? 24. Please list the major phases of the development process used. 25. What do you think is the most important phase(s) in developing a neural network system? (Please tick one) 26. How structured was your development methodology? 27. Do you think the development process is a complete methodology that involved all necessary stages? 28. Did the development process include constant refinement procedures? (i.e. did it support iterative procedures in every development stage?) 29. Were usability tests involved after the neural network system was built? 30. If yes, who were the people that were involved in the usability test? (Tick all that apply) 31. Was training offered to users after the neural network system was built? 32. Please tick the time required to complete the neural network system project? (Please tick one) 33. Please list up to 3 major benefits gained in implementing the neural network system. 34. How would you rate the overall effectiveness of your development process? 35. Thinking back, have you ever been involved in a neural network project that didn’t succeed as expected? 35a. If yes, what was the main reason for this? 36. Would you like a copy of the survey results? 36a. If yes, please provide us your email: (Note: Your email address will not be used in analysis) 37. Any comments you would like to add? |
Survey software
Leading survey software to help you turn data into decisions.
Research Edition
Intelligent market research surveys that uncover actionable insights.
Customer Experience
Experiences change the world. Deliver the best with our CX management software.
Workforce
Powerful insights to help you create the best employee experience.
