Apparatus and method for predicting the pleasantness-unpleasantness index of words using relative emotion similarity

• US 9,384,189 B2
• Filed: 10/21/2014
• Issued: 07/05/2016
• Est. Priority Date: 08/26/2014
• Status: Active Grant

• Current Assignee: Foundation of Soongsil University-Industry Cooperation
• Original Assignee: Foundation of Soongsil University-Industry Cooperation
• Inventors: Lee, Soo Won, Lee, Kang Bok
• Primary Examiner(s): ROBERTS, SHAUN A
• Application Number:
  US14/519,801
• Publication Number:
  US 20160062989A1
• Time in Patent Office: 623 Days
• Field of Search: 704/9
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3346 using probabilistic model
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06Q 10/04 Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202
• CPC Class Codes: G10L 25/63 for estimating an emotional state

Abstract

An apparatus and a method for predicting the pleasantness-unpleasantness index of words are disclosed. The disclosed apparatus includes: a computing unit configured to compute an emotion correlation between a word and one or more comparison word, compute emotion correlations between multiple reference words included in a reference word set and the one or more comparison word, compute multiple first absolute emotion similarity values between the word and the multiple reference words, and compute at least one second absolute emotion similarity value between a reference word and another reference word for all of the reference words included in the reference word set; and a prediction unit configured to predict the pleasantness-unpleasantness index of the word by using the multiple number of first absolute emotion similarity values, the at least one second absolute emotion similarity value, and a preset pleasantness-unpleasantness index of the multiple number of reference words.

Claims

1. An apparatus for predicting a pleasantness-unpleasantness index of a word using a computer, the apparatus comprising:

a computing unit configured to compute an emotion correlation between the word and one or more comparison word, compute emotion correlations between a plurality of reference words included in a reference word set and the one or more comparison word, compute a plurality of first absolute emotion similarity values between the word and the plurality of reference words, and compute at least one second absolute emotion similarity value between a reference word and another reference word for all of the plurality of reference words included in the reference word set; and

a prediction unit configured to predict the pleasantness-unpleasantness index of the word by using the plurality of first absolute emotion similarity values, the at least one second absolute emotion similarity value, and a preset pleasantness-unpleasantness index of the plurality of reference words, wherein

antonyms of the word are not included in the reference word set, and

the computing unit and the prediction unit are embodied on the computer.

2. The apparatus of claim 1, wherein the computing unit computes the emotion correlation between the word or the reference word and the comparison word by using a ratio between a probability of the word or the reference word and the comparison word appearing independently in a paragraph and a probability of the word or the reference word and the comparison word appearing together in a paragraph.

3. The apparatus of claim 1, wherein the computing unit computes a first vector and a plurality of second vectors, the first vector having the emotion correlation between the word and the one or more comparison words as an element, the second vectors having the emotion correlations between the plurality of reference words and the one or more comparison words as elements, and

the computing unit computes the plurality of first absolute emotion similarity values by using angles between the first vector and the plurality of second vectors, and computes the at least one second absolute emotion similarity value by using angles between the plurality of second vectors.

4. The apparatus of claim 3, wherein the computing unit computes a relative emotion similarity value between the word and each the plurality of reference words (a plurality of relative emotion similarity values) by using the plurality of first absolute emotion similarity values and the at least one second absolute emotion similarity value, and the prediction unit predicts the pleasantness-unpleasantness index of the word by using the plurality of relative emotion similarity values and the pleasantness-unpleasantness index of the plurality of reference words, and

an i-th relative emotion similarity value from among the plurality of relative emotion similarity values is computed using an i-th first absolute emotion similarity value between the word and an i-th reference word from among the plurality of reference words and second absolute emotion similarity values between the i-th reference word and the reference words of the reference word set other than the i-th reference word.

5. The apparatus of claim 4, wherein the computing unit computes the i-th relative emotion similarity value by using an equation shown below:

6. The apparatus of claim 4, wherein the prediction unit predicts the pleasantness-unpleasantness index of the word by using an equation shown below:

7. A method for predicting a pleasantness-unpleasantness index of a word using a computer, the method comprising:

computing, by the computer, an emotion correlation between the word and one or more comparison word;

computing, by the computer, emotion correlations between a plurality of reference words included in a reference word set and the one or more comparison word;

computing, by the computer, a plurality of first absolute emotion similarity values between the word and the plurality of reference words;

computing, by the computer, at least one second absolute emotion similarity value between a reference word and another reference word for all of the plurality of reference words included in the reference word set; and

predicting, by the computer, the pleasantness-unpleasantness index of the word by using the plurality of first absolute emotion similarity values, the at least one second absolute emotion similarity value, and a preset pleasantness-unpleasantness index of the plurality of reference words,

wherein antonyms of the word are not included in the reference word set.

8. A non-transitory computer-readable recorded medium tangibly embodying a program of instructions for performing the method of claim 7.