Method and server for predicting a query-completion suggestion for a partial user-entered query

A method of training a machine learning algorithm (MLA) for determining a query-completion suggestion for a partial user-entered query, the partial user-entered query being part of a potential user query, the method executable by a server, the method comprising:

• receiving, by the server, a plurality of past user queries;
• parsing, by the server, the plurality of past user queries into a plurality of n-grams, each one of the plurality of n-grams being associated with at least one respective past user query and respective n-gram features, the n-gram features being indicative of a pair-based co-occurrence of n-grams from each possible pair of n-grams of the plurality of n-grams in the plurality of past user queries;
• and
• for a given n-gram of a given past user query:
• selecting, by the server, at least one candidate n-gram from the plurality of n-grams based on the pair-based co-occurrence of each one of the plurality of n-grams and the given n-gram in the plurality of past user queries;
• generating, by the server, respective feature vectors for the given n-gram and the at least one candidate n-gram based on the respective n-gram features;
• generating, by the server, a training set for the given n-gram comprising an input portion and a label portion,
• the input portion being based on the given n-gram, the at least one candidate n-gram and the respective feature vectors,
• the label portion being based on the given past user query and the at least one candidate n-gram, the label portion being indicative of a group-based co-occurrence of a group of n-grams in the given past user query, the group of n-grams comprising (i) at least one of the at least one candidate n-gram and (ii) the given n-gram, and
• training, by the server, the MLA based on the training set to determine, during its in-use phase, a predicted group-based co-occurrence of at least one in-use candidate n-gram and at least one in-use n-gram in the potential user query, the at least one in-use n-gram being the partial user-entered query, the predicted group-based co-occurrence for determining which of the at least one in-use candidate n-gram is to be used as the respective query-completion suggestion.