User input prediction

A method of inferring term boundaries in an input sequence comprising a contiguous sequence of characters, the method comprising:

• generating one or more candidates from the input sequence, wherein each candidate comprises two terms separated by one or more term boundaries;
• searching, using a processor, for one or more terms of each candidate in stored sequences of terms, each sequence of terms having a corresponding probability of occurring;
• and
• assigning a probability to each candidate based on the search.

A system comprising:

• a processor;
• memory storing instructions that, when executed by the processor, configure the processor to:
• generate one or more candidates from an input sequence, the input sequence comprising a contiguous sequence of characters, wherein each candidate comprises two or more terms separated by one or more term boundaries;
• wherein the processor generates one or more candidates from the input sequence by being configured to:
• convert the input sequence into a graph comprising one or more paths;
• and
• insert one or more term boundaries into the graph, between two neighbouring characters of the input sequence,
• wherein the one or more paths of the graph correspond to the one or more candidates.