You contact me about this project 10 days ago. The following was your message
I developed a suffix tree (slight improvement over a trie) about two years ago, written in Java. I used it on a passage of text to count the frequencies that a given substring occurred in the text. I no longer have the source code, but I would be willing to write it from scratch. I would plan on designing it as a node-based tree. I originally wrote the suffix tree in Java, but would be willing to code in .NET as well, as I use .NET more often nowadays. Your choice. PM me if you have questions
------------- My Original Description ----------------------------
This project is not for commercial use but for academic proof of concept purposes only.
Basically the program that I am looking for will be able to predict the next few words which the user will be about to write. This program must be able to store SMS phrases both incoming and outgoing. Upon receiving a message the program must also predict the most probable reply based on similar messages and give the user a number of options which he/she could choose from. The program must learn the trends of the particular user and thus be able to predict the most likely reply based on past user behaviour.
For the time being this program will not be deployed on a mobile phone. The program will have 2 textboxes which will simulate send and received of a mobile phone. 1 text box for receiving and one for sending.
When the user wants to send a message, he/she will type the message in the sending textbox. As soon as the user starts typing the first few letters or a word the program will prompt a possible next predicted word(s) or phrase with a highest frequency. There should be a key, which when hit that shows the second highest frequent possible match.
Unless you have a better suggestion the best way for such thing is Ternary-Search-Tree. If the mobile will be switch off(as a concept in our case the application will be closed) data have to be saved from the trie to a file that is why I requested that messages/phrases will be loaded from a file.
Thus messaging is not between 2 applications over TCP/IP, but between two windows (let say two textboxes) in the same program interface.
Text prediction as implemented by Yahoo! (I can't send you an image) in the website's search field. The program similarly enable the prediction of the next few words whilst compiling instant messages on mobile phones.
The program must predict most likely replies to a message received based on trie and on past history.
If you would like further details please feel free to contact me. Can you do this, I am waiting for your reply.
Also could you tell me when you can start and how long.
Kindly can you BID again here,
Thank & Regards