Xiao Ding-Dong robot, decision tree, iask, guess 20 (http://y.20q.net/anon)
In the morning, a friend of mine told me that the robot was stupid and not fun. This is why I realized that the robot has not been updated for about half a year, but I hope he can grow up quickly.
About last year, chatbots became popular on the Internet.
Xiaoqiang and other robots, including robots provided by QQ, are not positioned as "chatting", but as personal information tools. For example, you can query the region, IP address, zip code, and weather conditions of the phone number.
And other personal services. Obviously, because of the single communication mode between users and the backend (chat only), the advantages of this service are not obvious. That is why I had the idea of making robots into a system similar to an expert system when I was a little robot.
It is very difficult to make Xiao Dingdong robot into an expert system.
For Xiao Ding-Dong robot, the Chinese word segmentation and automatic classification, coupled with a inference engine, seem to be able to upgrade Xiao Ding-Dong robot's IQ to a higher level. The required technologies are as follows:
First, let the robot understand the user input conversation;
Second, we need to classify conversations and then reason based on our own knowledge base to respond accordingly.
This is the original intention of studying Chinese word segmentation. Chinese word segmentation is the basis for automatic processing by many Chinese computers, such as search engines, automatic summarization, and so on. Recently, many netizens wrote to me to ask related questions. I feel this part is really a hot topic.
After the Chinese word segmentation, You need to classify it. This part of Sina iask is a typical application. I think this implementation is not difficult;
Recently I have seen a question 20. I think this is a good application of decision tree learning. Maybe I should consider implementing a similar application now, prepare for the full improvement of Xiao Dingdong's IQ.
Let's start with building a decision tree.
Related connections:
How happy
Chinese Word Segmentation
Sina iask
Guess 20