UPDATE: We got operational funding from NTU!
Moving on from this point onwards, we will need
TO DO LIST:
Get the basic essential and easy features up, especially on FAQs, so that it can be concluded and we can proceed on to the harder parts.- Set up the basic essential data storage, through the python-telegram-bot API "persistence" so that we can build further into this and to actually store and conduct multiple transactions. Will need CS help on this since I am not very well-versed into this.
Do a bit of sudocode in the sense of how we want the transaction (order, and order status) to be within the bot. The initial set-up and mechanisms needs to be discussed ASAP so that we can have a bit of understanding what should go into the bot.Suggestions would also need to be done in the sense that it will have to include polls (for people to poll) and for customer feedback too (through that as well).Once all of this is done right, we move on to the more difficult machine-learning parts.
DIFFICULT:
Need to figure out machine learning (kNN, linear regression, yada yada) in the sense that using the "How are you feeling" function, we can pick up certain tones and provide recommendations to the end-user.- We also need to provide recommendations to the order-list such that over time, the bot understands and can interpret how the end-user is feeling through phrases, sentences and words so that key recommendations can be provided. It is important to note that the same item should not be given twice, and that the item should be pre-approved by the end-user (unless it is a mystery).
- Figure out simply the long-term objectives of this bot basically, and how are we going to integrate this bot into bigger things or sponsorships.
A bit tired writing this so feel free to add in more to-do list if you feel like it haha
Bot source code has been made public upon request.