The English sentences have patterns; the patterns can be grouped according to the meaning they express. Because different patterns can express the same meaning, people can ask the same question in different ways. It's hard for a chatbot developer to anticipate all the possible ways to ask the question. By grouping the patterns that express the same meaning, we can specify one most commonly used pattern as Final Pattern; all other patterns in the group will be converted into that final pattern. That way the chatbot developer will only need to write an answer to the question using the final pattern. If the customer uses a different pattern to ask the question, it can be recognized and answered. For example, we have a group of patterns such as "what sorts of ... ", "what sort of ... ", "what types of ...", "what type of ...", "what kinds of ...", "what kind of ..."; if we decide to use "what kind of ..." as final pattern for the group, the chatbot developer will only need to answer the question "what kind of (services do you have)" and don't need to worry about whether the customer would use "what sort of …" or "what type of ..." That is how Final Patterns approach improves Natural Language Understanding and shorten the time in creating a chatbot.
The success of a chatbot largely depends on how well the chatbot understands the customer. The Final Patterns, once utilized in a chatbot, will make it understand the customers better. Through years of study of English language patterns, Dr. Ron Lee has identified over 100 final patterns, and put them into a computer programming script. The Final Patterns Script can shorten the time in the development of chatbots, or any conversational systems. The script has been tested over years with Chatbot Tutor. Because of the use of the Final Patterns Script, Tutor has achieved great success in the annual Loebner prize contest.
Starting in 2019, we provide services to help people develop chatbots for various purposes. Please send inquiry to email@example.com.