SYSTEMS AND METHODS FOR IDENTIFYING AN INTENT OF A USER QUERY
Systems and methods are described herein for identifying an intent of a user query. A media guidance application may receive a query from a user using any suitable method, including text input and voice/speech recognition software. The media guidance application may extract a plurality of n-grams from the query. The media guidance application may calculate, for each of the plurality of n-grams, a mutual information value with each of a plurality of potential intents. The mutual information value may indicate a measure of dependence between one of the plurality of n-grams and one of the plurality of potential intents. The media guidance application may select a subset of the plurality of n-grams based on the calculated mutual information values and input the subset of the plurality of n-grams into a probabilistic classifier, such as a Naïve Bayes Classifier, in order to identify the intent of the query.
1 . A method for identifying an intent of a user query, comprising:
receiving a query from a user;
extracting a plurality of n-grams from the query;
calculating, for each of the plurality of n-grams, a mutual information value with each of a plurality of potential intents;
selecting a subset of the plurality of n-grams based on the calculated mutual information values; and
identifying the intent of the query by inputting the subset of the plurality of n-grams into a probabilistic classifier.
2 . The method of claim 1 , wherein the probabilistic classifier identifies the intent of the query based on conditional probabilities of the subset of the plurality of n-grams being related to each of the plurality of potential intents.
3 . The method of claim 1 , wherein the probabilistic classifier is a Naïve Bayes classifier.
4 . The method of claim 1 , wherein the calculated mutual information values indicate a measure of dependence between one of the plurality of n-grams and one of the plurality of potential intents.
5 . The method of claim 1 , wherein each n-gram of the subset is associated with a mutual information value with at least one of the plurality of potential intents that exceeds a threshold value.
6 . The method of claim 1 , wherein selecting the subset of the plurality of n-grams based on the calculated mutual information values comprises removing an n-gram from the subset associated with calculated mutual information values that do not exceed a threshold value.
7 . The method of claim 1 , wherein selecting the subset of the plurality of n-grams based on the calculated mutual information values comprises removing an n-gram from the subset identified in a list of stop words.
8 . The method of claim 1 , further comprising:
identifying a dominating n-gram of the plurality of n-grams, wherein the dominating n-gram is associated with a mutual information value with a dominating intent of the plurality of potential intents that exceeds a threshold value; and
wherein identifying the intent of the query comprises increasing a probability that the intent of the query is the dominating intent.
9 . The method of claim 8 , wherein identifying the intent of the query further comprises decreasing a probability that the intent of the query is an intent other than the dominating intent.
10 . The method of claim 1 , wherein receiving the query from the user comprises receiving the query using one of text input or speech recognition.
11 . A system for identifying an intent of a user query, comprising:
control circuitry configured to:
receive a query from a user;
extract a plurality of n-grams from the query;
calculate, for each of the plurality of n-grams, a mutual information value with each of a plurality of potential intents;
select a subset of the plurality of n-grams based on the calculated mutual information values; and
identify the intent of the query by inputting the subset of the plurality of n-grams into a probabilistic classifier.
12 . The system of claim 11 , wherein the probabilistic classifier identifies the intent of the query based on conditional probabilities of the subset of the plurality of n-grams being related to each of the plurality of potential intents.
13 . The system of claim 11 , wherein the probabilistic classifier is a Naïve Bayes classifier.
14 . The system of claim 11 , wherein the calculated mutual information values indicate a measure of dependence between one of the plurality of n-grams and one of the plurality of potential intents.
15 . The system of claim 11 , wherein each n-gram of the subset is associated with a mutual information value with at least one of the plurality of potential intents that exceeds a threshold value.
16 . The system of claim 11 , wherein the control circuitry is configured to select the subset of the plurality of n-grams based on the calculated mutual information values by removing an n-gram from the subset associated with calculated mutual information values that do not exceed a threshold value.
17 . The system of claim 11 , wherein the control circuitry is configured to select the subset of the plurality of n-grams based on the calculated mutual information values by removing an n-gram from the subset identified in a list of stop words.
18 . The system of claim 11 , wherein the control circuitry is further configured to:
identify a dominating n-gram of the plurality of n-grams, wherein the dominating n-gram is associated with a mutual information value with a dominating intent of the plurality of potential intents that exceeds a threshold value; and
wherein the control circuitry is configured to identify the intent of the query by increasing a probability that the intent of the query is the dominating intent.
19 . The system of claim 18 , wherein the control circuitry is configured to identify the intent of the query by decreasing a probability that the intent of the query is an intent other than the dominating intent.
20 . The system of claim 11 , wherein the control circuitry is configured to receive the query from the user by receiving the query using one of text input or speech recognition.
21 - 50 . (canceled)