BotMakr uses a series of proprietary algorithms for matching a user's query with a bot's corpus. Here is a brief description of each algorithm:
Matches phrases using a natural language matching algorithm.
Same as N1 but uses additional context for matching. This algorithm is better in cases where the user enters long queries, especially when they contain frequent misspellings.
Matches on how closely the user's query is contained other corpus phrases, and vice-versa.
Matches on audio deltas between the user's query and the bot's corpus phrases. Good for when the user tends to enter short phrases, especially when they contain frequent misspellings.
Matches on lexical deltas between the user's query and the bot's corpus phrases. Good for when the user tends to enter short phrases
A combination of L1 and L2.
- If you give every algorithm a weight of 0.0, your bot will default to using N1.
- N1 and N2 are intended to be either/or, i.e. there is no point to using both. If you use one, it is suggested that you set the other to 0.
- Due to performance reasons, L1 and L2 are only used for bots that have a relatively small corpus. If you use L1 and L2, it is suggested that you also use one of the other algorithms as a supplement in case your bot's corpus grows beyond the L1 and L2 threshold.