To solve this, we combined Bing Search API with a web screenshot API to save a local image. While text-based SMS send/return was fairly straightforward, we ran into several challenges when sending images of the web page as a response. We used multiple artificial intelligence based natural language processing algorithms to analyze the key words of queries and sentiment analysis of the top articles to display only accurate and positive web pages to the user. For the web images, we used Bing Search to query the users search parameters and an web screenshot API to save this image, where the returned PNG image is saved locally and sent back to the user via MMS image. We created our own custom natural language processing chatbot to prompt the user with followup questions. Based on user requests and parameters, various search methods are used to collect the information and our processing algorithm formats the return JSON into a return SMS based on user preference. All information collected from the user is processed by our text parsing algorithms to find the best way to communicate with the user. We used the Twilio API to receive and send SMS messages from the user through our express node.js script running on the Azure backend.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |