|
The purpose of this project is to create a program capable of reading
sentences and gathering information from them so that, when asked a question
later, the program can search its knowledge database and formulate an
answer. This requires that some ambiguities in the way sentences are
structured must be resolved so the program has a better chance of
understanding what is said. This would allow otherwise computer-illiterate
people to use the computer to find information. The program currently does
not support phrases or clauses, and it doesn't interpret what is being read
to make inferences, nor does it support nouns more than one word long. The
program breaks down the sentences it receives, determines the sentence
pattern in use, and then determines the part of speech of each word in a
sentence. It then records this information in its knowledge database. When a
user asks it a question, it refers back to this database to find the answer.
Within its limits, the program correctly reads sentences, resolving
ambiguous structures it encounters. The limits of the program are easily
extensible to include the ability to read phrases and clauses, draw
conclusions from what's being read, understand words with multiple meanings,
and combine sentences that are output to seem more human.
|