This task will involve summarizing emails, email reply predictions and email groupings based on activities of the user ( Clustering emails using word extraction)
The three functions above should be implemented with any algorithm like Bayesian classifications, SVM or any other using Perl or c/c++ programming language.
1. Email summarizer: should be able to summarise email message contents. For example if we have an email with 5 sentences, the algorithm should be able to summarise and pick up the important message and summarise it to 1 or 2 sentences. Just picking up important or repeated words and summarise it.
2. Email Groupings: This is just to group email message based on users' activities. Not folders. Grouping based on what the users do or contents of the mail
3. Email reply predictions: This is to have an algorithm that can predict that the email message that we receive require a reply. May be based on time line or thread or something else.
I will be glad if this could be completed within 4 weeks.
(CAN YOU USE WEKA OR BOGOFILTER OR POPFILE or something different).