Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Can Social Comments Contribute to Estimate Impression of Music Video Clips?

140 views

Published on

CollabTech2018で発表した際のプレゼンスライドになります。

Published in: Data & Analytics
  • Be the first to comment

  • Be the first to like this

Can Social Comments Contribute to Estimate Impression of Music Video Clips?

  1. 1. Can Social Comments Contribute to Estimate Impression of Music Video Clips? Shunki Tsuchiya (Meiji University) Naoki Ono, Satoshi Nakamura, and Takehiro Yamamoto
  2. 2. Outline of our study • We want to be able to search music video clips (MV) by impressions. Exciting Music video clips • We must estimate the impression of MV using social comments. Exciting? Happy? Cute? Cheerful? Cool?
  3. 3. Contributions 1. We generated the impression dataset of chorus part of 500 MV in three media types. • music only • video picture only • combined 2. It is better to use proper parts-of-speech in social comments depending on each media/impression type.
  4. 4. Background
  5. 5. • Anyone can publish their own MV by the spread of Consumer-generated media websites and DTM software. Background 1 The number of MV on the Web has dramatically increased.
  6. 6. Background 2 A standard method of searching for MV is to input keywords. Beatles Let It Be If a user doesn’t know the information, it is not easy to find the target MV. ????
  7. 7. Background 2 A standard method of searching for MV is to input keywords. Beatles Let It Be If a user doesn’t know the information, it is not easy to find the target MV. ???? In this search method, a user need to know the information in advance.
  8. 8. When you listen to music… How do you search for that BGM? The name of song The name of artist
  9. 9. When you listen to music… You will receive some impression such as “fierce” or “cheerful”.
  10. 10. The impression-based search I have happy things We research ambiguous searches based on the user’s subjective impressions. ex) exciting, cheerful, cute, sorrowful I’m feel like crying, so I want to watch and listen to sad MV. I feel happy, so I want to watch and listen to cheerful MV.
  11. 11. Impression information There are few impression tags attached. Percentage of impression tag 5%[Yamamoto2009] 14%[Hu2007] It is too difficult for people to evaluate the impressions of all MV. It is necessary that impression information is given to the MV.
  12. 12. Impression estimation We have to evaluate and provide subjective impressions on individual MV in advance. Exciting? Happy? Cute? Cheerful? Cool? If it becomes possible to estimate the impression and available by API, it can be used for creating new service using MV
  13. 13. Impression estimation We have to evaluate and provide subjective impressions on individual MV in advance. Exciting? Happy? Cute? Cheerful? Cool? If it becomes possible to estimate the impression and available by API, it can be used for creating new service using MV We expect that impressions are used not only for searching but also various things.
  14. 14. Related work 1 Studies on estimates of impressions of MV Music Movies • Understanding Affective Content of music video through learned representations. [Acar 2014] • Multimedia content analysis for emotional characterization of music video clips. [Ashkan 2013] Music video clips These studies estimated impressions by using these features.
  15. 15. Related work 1 Studies on estimates of impressions of MV Music Movies • Understanding Affective Content of music video through learned representations. [Acar 2014] • Multimedia content analysis for emotional characterization of music video clips. [Ashkan 2013] Music video clips These studies estimated impressions by using these features. Features of music and movies are mechanical, where no human emotions are reflected
  16. 16. Related work 2 Studies on estimates of impressions of… • music using lyrics • movies using viewer’s expressions • Lyric text mining in music mood classification. [Hu 2009] • Video indexing and recommendation based on affective analysis of viewers. [Sicheng 2009] We also use subjective features. The estimation accuracy is improved by using subjective features.
  17. 17. Social comments 1 http://www.nicovideo.jp/watch/sm13252011 Nico Nico Douga and BiliBili has a function to provide comments in real time to the video. 72 million users35 million users
  18. 18. Social comments 2 • Users can use this function to communicate with others. • We consider that these comments express impressions that users directly in real time. We estimate impressions of MV from the comments used for this communication
  19. 19. Study of Yamamoto [2013] • Yamamoto’s study used only adjectives in comments to estimate impression. → We also target other parts-of-speech. • Yamamoto’s study used the whole of a MV for estimation. →We estimate the impressions of only the chorus part. Some MV Start End Verse 1st Chorus Bridge 2nd Chorus Exciting Happy Painful Happy
  20. 20. Media types We considered media types that are music only, video picture only, and combined. The impression received from music only, video picture only and combined was different.
  21. 21. Purpose We examine the degree of impression estimation accuracy of MV using social comments. • Estimate impressions by using the part-of-speech in comments • Estimate impressions of the chorus part • Consider differences by media type
  22. 22. Impression dataset
  23. 23. Generating the dataset • Target 500 MV ( ) ( ( 30s of the chorus part“VOCALOID” ReflaiD [Goto2006] • Divides a MV into 3 media types
  24. 24. Impression types Impression names Adjectives representing impressions C1 (exciting) Exciting, bustling, proudly, & dignified C2 (cheerful) Cheerful, happy, hilarious, & comfortable C3 (painful) Painful, gloomy, bittersweet, & sorrowful C4 (fierce) Fierce, aggressive, emotional, & active C5 (humorous) Humorous, funny, strange, & capricious C6 (cute) Cute, lovely, awesome, &tiny Valence Bright feelings, fun & dark feelings, sad Arousal Aggressive, bullish & passive, bearish
  25. 25. Impression types Impression names Adjectives representing impressions C1 (exciting) Exciting, bustling, proudly, & dignified C2 (cheerful) Cheerful, happy, hilarious, & comfortable C3 (painful) Painful, gloomy, bittersweet, & sorrowful C4 (fierce) Fierce, aggressive, emotional, & active C5 (humorous) Humorous, funny, strange, & capricious C6 (cute) Cute, lovely, awesome, &tiny Valence Bright feelings, fun & dark feelings, sad Arousal Aggressive, bullish & passive, bearish C1 (exciting) Exciting, bustling, proudly, & dignified C2 (cheerful) Cheerful, happy, hilarious, & comfortable C3 (painful) Painful, gloomy, bittersweet, & sorrowful C4 (fierce) Fierce, aggressive, emotional, & active C5 (humorous) Humorous, funny, strange, & capricious Five impressions used in MIREX [Hu 2008]
  26. 26. Impression types Impression names Adjectives representing impressions C1 (exciting) Exciting, bustling, proudly, & dignified C2 (cheerful) Cheerful, happy, hilarious, & comfortable C3 (painful) Painful, gloomy, bittersweet, & sorrowful C4 (fierce) Fierce, aggressive, emotional, & active C5 (humorous) Humorous, funny, strange, & capricious C6 (cute) Cute, lovely, awesome, &tiny Valence Bright feelings, fun & dark feelings, sad Arousal Aggressive, bullish & passive, bearish C6 (cute) Cute, lovely, awesome, &tiny Many tags that are “cute” are attached on Nico Nico Douga [Yamamoto 2013].
  27. 27. Impression types Impression names Adjectives representing impressions C1 (exciting) Exciting, bustling, proudly, & dignified C2 (cheerful) Cheerful, happy, hilarious, & comfortable C3 (painful) Painful, gloomy, bittersweet, & sorrowful C4 (fierce) Fierce, aggressive, emotional, & active C5 (humorous) Humorous, funny, strange, & capricious C6 (cute) Cute, lovely, awesome, &tiny Valence Bright feelings, fun & dark feelings, sad Arousal Aggressive, bullish & passive, bearish Two impressions called valence-arousal space proposed by Russell [Russell 1980]. Valence Bright feelings, fun & dark feelings, sad Arousal Aggressive, bullish & passive, bearish
  28. 28. Method of impression evaluation • One of the media for 30s • Random regardless of media type • Answered each impression with a 5 rank Likert scale (-2, -1, 0, +1, +2) 3 subjects evaluated one of the media.
  29. 29. Impression dataset Some MV C1 C2 C3 C4 C5 C6 V A Combined -1.3 -2.0 -0.3 0.0 1.7 -2.0 -0.7 -0.7 Music only -1.7 -2.0 2.0 0.0 -1.7 -2.0 0.3 -1.7 Video picture only 0.3 1.3 -0.3 -0.7 -0.7 1.7 -0.3 1.7 We used the average value of 3 subjects as the impression evaluation value for each media/impression type.
  30. 30. Evaluation experiment
  31. 31. Outline of evaluation experiment We tested and verified by using SVMs whether impression having an evaluation of more than a certain value could be mechanically estimated. 1. Method of impression estimation 2. Number of MV in each evaluation group 3. Collecting and extracting social comments 4. Generation of bag-of-words for MV 5. Method of bag-of-words generation Details
  32. 32. Method of impression estimation 3 media 8 impressions = 24 pattern Low evaluation group High evaluation group Training data Test data We verified high evaluation group could be mechanically estimated by using SVMs.
  33. 33. High group C1 C2 C3 C4 C5 C6 V A Combined 76 105 87 54 83 104 101 150 Music only 133 127 46 69 49 73 124 178 Video picture only 21 50 142 49 81 78 57 111 Number of MV in each group Low group C1 C2 C3 C4 C5 C6 V A Combined 105 169 191 209 178 215 62 94 Music only 65 92 232 195 180 209 61 43 Video picture only 252 272 165 247 207 234 96 155
  34. 34. High group C1 C2 C3 C4 C5 C6 V A Combined 76 105 87 54 83 104 62 94 Music only 65 92 46 69 49 73 61 43 Video picture only 21 50 142 49 81 78 57 111 Number of MV in each group Low group C1 C2 C3 C4 C5 C6 V A Combined 76 105 87 54 83 104 62 94 Music only 65 92 46 69 49 73 61 43 Video picture only 21 50 142 49 81 78 57 111
  35. 35. Collecting and extracting social comments • We collected all comments (860,455) for 500 MV using Nico Nico Douga API. • We extracted comments (132,036) posted in the chorus part. Some MV Start End Verse 1st Chorus Bridge 2nd Chorus
  36. 36. Generation of bag-of-words for MV / / / Comments of a MV = “Miku is cute.” “The melody is cool.” “Miku is good”= “Miku / cute.” “Melody / cool.” “Miku / good” Miku Cute Melody Cool Good 2 1 1 1 1 Miku Melody 2 1 Cute Cool Good 1 1 1 All parts-of-speech Nouns Adjectives
  37. 37. Methods of bag-of-words generation Method names Parts-of- speech used All method All parts-of- speech All2 method Nouns, Verbs, Adjectives, Adverbs Noun method Nouns Verb method Verbs Adj method Adjectives Adv method Adverbs Method names Parts-of- speech used Noun-Verb method Nouns, Verbs Noun-Adj method Nouns, Adjectives Noun-Adv method Nouns, Adverbs Verb-Adj method Verbs, Adjectives Verb-Adv method Verbs, Adverbs Adj-Adv method Adjectives, Adverbs
  38. 38. Results
  39. 39. Noun C1 C2 C3 C4 C5 C6 V A Comb 0.575 0.720 0.644 0.653 0.704 0.680 0.646 0.652 Music 0.698 0.606 0.528 0.621 0.721 0.661 0.708 0.650 Video 0.700 0.640 0.608 0.600 0.620 0.688 0.552 0.641 Noun vs Verb vs Adj vs Adv Verb C1 C2 C3 C4 C5 C6 V A Comb 0.667 0.627 0.440 0.544 0.642 0.714 0.575 0.574 Music 0.615 0.622 0.133 0.658 0.587 0.500 0.600 0.551 Video 0.588 0.549 0.606 0.517 0.584 0.573 0.508 0.654
  40. 40. Adj C1 C2 C3 C4 C5 C6 V A Comb 0.733 0.869 0.710 0.750 0.667 0.838 0.650 0.842 Music 0.667 0.635 0.595 0.667 0.581 0.775 0.706 0.733 Video 0.714 0.736 0.733 0.759 0.536 0.829 0.603 0.850 Nous vs Verb vs Adj vs Adv Adv C1 C2 C3 C4 C5 C6 V A Comb 0.618 0.586 0.522 0.576 0.520 0.481 0.556 0.603 Music 0.679 0.600 0.580 0.537 0.545 0.481 0.642 0.538 Video 0.879 0.759 0.211 0.632 0.519 0.451 0.777 0.805
  41. 41. Nous vs Verb vs Adj vs Adv • Users most often expressed impressions using adjective. • We consider that the adjective words have features for impression, and noun, verb, and adverb didn’t have. Adjective > Noun, Verb, Adverb
  42. 42. N-V C1 C2 C3 C4 C5 C6 V A Comb 0.687 0.699 0.648 0.620 0.681 0.714 0.661 0.636 Music 0.683 0.580 0.489 0.642 0.689 0.672 0.729 0.658 Video 0.881 0.760 0.308 0.614 0.595 0.639 0.805 0.859 Method without including Adj N-Av C1 C2 C3 C4 C5 C6 V A Comb 0.592 0.714 0.644 0.654 0.722 0.673 0.656 0.649 Music 0.672 0.589 0.538 0.621 0.711 0.661 0.694 0.632 Video 0.879 0.763 0.372 0.636 0.622 0.683 0.805 0.852 V-Av C1 C2 C3 C4 C5 C6 V A Comb 0.667 0.568 0.535 0.531 0.657 0.630 0.600 0.660 Music 0.667 0.560 0.458 0.566 0.587 0.513 0.589 0.581 Video 0.882 0.729 0.250 0.622 0.488 0.529 0.724 0.814 Noun-Verb Noun-Adv Verb-Adv
  43. 43. Aj-Av C1 C2 C3 C4 C5 C6 V A Comb 0.700 0.837 0.679 0.690 0.681 0.848 0.695 0.844 Music 0.733 0.646 0.581 0.634 0.683 0.743 0.667 0.718 Video 0.911 0.765 0.477 0.653 0.622 0.757 0.840 0.884 V-Aj C1 C2 C3 C4 C5 C6 V A Comb 0.781 0.811 0.711 0.684 0.667 0.856 0.652 0.784 Music 0.692 0.627 0.520 0.714 0.682 0.740 0.673 0.707 Video 0.921 0.734 0.400 0.734 0.511 0.764 0.779 0.871 N-Aj C1 C2 C3 C4 C5 C6 V A Comb 0.662 0.854 0.690 0.780 0.750 0.778 0.694 0.80 Music 0.754 0.644 0.612 0.750 0.707 0.772 0.740 0.806 Video 0.888 0.792 0.409 0.706 0.657 0.768 0.821 0.874 Method including Adj Noun-Adj Verb-Adj Adj-Adv
  44. 44. Method of combining > Method of combining part-of-speech Method using only one part-of-speech N-V C1 C2 C3 C4 C5 C6 V A Comb 0.687 0.699 0.648 0.620 0.681 0.714 0.661 0.636 Music 0.683 0.580 0.489 0.642 0.689 0.672 0.729 0.658 Video 0.881 0.760 0.308 0.614 0.595 0.639 0.805 0.859 Noun C1 C2 C3 C4 C5 C6 V A Comb 0.575 0.720 0.644 0.653 0.704 0.680 0.646 0.652 Music 0.698 0.606 0.528 0.621 0.721 0.661 0.708 0.650 Video 0.700 0.640 0.608 0.600 0.620 0.688 0.552 0.641 Verb C1 C2 C3 C4 C5 C6 V A Comb 0.667 0.627 0.440 0.544 0.642 0.714 0.575 0.574 Music 0.615 0.622 0.133 0.658 0.587 0.500 0.600 0.551 Video 0.588 0.549 0.606 0.517 0.584 0.573 0.508 0.654 Noun-Verb Noun Verb The combination of parts-of-speech improved the accuracy of estimation.
  45. 45. Aj- Av C1 C2 C3 C4 C5 C6 V A Comb 0.700 0.837 0.679 0.690 0.681 0.848 0.695 0.844 Music 0.733 0.646 0.581 0.634 0.683 0.743 0.667 0.718 Video 0.911 0.765 0.477 0.653 0.622 0.757 0.840 0.884 N-Aj C1 C2 C3 C4 C5 C6 V A Comb 0.662 0.854 0.690 0.780 0.750 0.778 0.694 0.80 Music 0.754 0.644 0.612 0.750 0.707 0.772 0.740 0.806 Video 0.888 0.792 0.409 0.706 0.657 0.768 0.821 0.874 Method of combining > Method of combining part-of-speech Method using only one part-of-speech Adj Noun-Adj Adj-Adv The combination of parts-of-speech improved the accuracy of estimation. Adj C1 C2 C3 C4 C5 C6 V A Comb 0.733 0.869 0.710 0.750 0.667 0.838 0.650 0.842 Music 0.667 0.635 0.595 0.667 0.581 0.775 0.706 0.733 Video 0.714 0.736 0.733 0.759 0.536 0.829 0.603 0.850
  46. 46. C1 C2 C3 C4 C5 C6 V A Ave Comb V-Adj 0.781 Adj 0.869 All 0.713 N-Adj 0.780 N-Adj 0.750 V-Adj 0.856 All 0.783 Aj-Av 0.844 0.797 Music N-Adj 0.754 All 0.671 N-Adj 0.612 N-Adj 0.750 All2 0.725 All2 0.787 N-Adj 0.740 N-Adj 0.806 0.730 Video V-Adj 0.921 N-Adj 0.792 All 0.752 Adj 0.759 N-Adj 0.657 Adj 0.829 Aj-Av 0.840 Aj-Av 0.884 0.804 Ave 0.819 0.777 0.692 0.763 0.711 0.824 0.788 0.845 0.777 The highest value and its method
  47. 47. C1 C2 C3 C4 C5 C6 V A Ave Comb V-Adj 0.781 Adj 0.869 All 0.713 N-Adj 0.780 N-Adj 0.750 V-Adj 0.856 All 0.783 Aj-Av 0.844 0.797 Music N-Adj 0.754 All 0.671 N-Adj 0.612 N-Adj 0.750 All2 0.725 All2 0.787 N-Adj 0.740 N-Adj 0.806 0.730 Video V-Adj 0.921 N-Adj 0.792 All 0.752 Adj 0.759 N-Adj 0.657 Adj 0.829 Aj-Av 0.840 Aj-Av 0.884 0.804 Ave 0.819 0.777 0.692 0.763 0.711 0.824 0.788 0.845 0.777 The highest value and its method
  48. 48. C1 C2 C3 C4 C5 C6 V A Ave Comb V-Adj 0.781 Adj 0.869 All 0.713 N-Adj 0.780 N-Adj 0.750 V-Adj 0.856 All 0.783 Aj-Av 0.844 0.797 Music N-Adj 0.754 All 0.671 N-Adj 0.612 N-Adj 0.750 All2 0.725 All2 0.787 N-Adj 0.740 N-Adj 0.806 0.730 Video V-Adj 0.921 N-Adj 0.792 All 0.752 Adj 0.759 N-Adj 0.657 Adj 0.829 Aj-Av 0.840 Aj-Av 0.884 0.804 Ave 0.819 0.777 0.692 0.763 0.711 0.824 0.788 0.845 0.777 The highest value and its method It is better to use proper parts-of- speech depending on each types.
  49. 49. C1 C2 C3 C4 C5 C6 V A Ave Comb V-Adj 0.781 Adj 0.869 All 0.713 N-Adj 0.780 N-Adj 0.750 V-Adj 0.856 All 0.783 Aj-Av 0.844 0.797 Music N-Adj 0.754 All 0.671 N-Adj 0.612 N-Adj 0.750 All2 0.725 All2 0.787 N-Adj 0.740 N-Adj 0.806 0.730 Video V-Adj 0.921 N-Adj 0.792 All 0.752 Adj 0.759 N-Adj 0.657 Adj 0.829 Aj-Av 0.840 Aj-Av 0.884 0.804 Ave 0.819 0.777 0.692 0.763 0.711 0.824 0.788 0.845 0.777 The highest value and its method
  50. 50. C1 C2 C3 C4 C5 C6 V A Ave Comb V-Adj 0.781 Adj 0.869 All 0.713 N-Adj 0.780 N-Adj 0.750 V-Adj 0.856 All 0.783 Aj-Av 0.844 0.797 Music N-Adj 0.754 All 0.671 N-Adj 0.612 N-Adj 0.750 All2 0.725 All2 0.787 N-Adj 0.740 N-Adj 0.806 0.730 Video V-Adj 0.921 N-Adj 0.792 All 0.752 Adj 0.759 N-Adj 0.657 Adj 0.829 Aj-Av 0.840 Aj-Av 0.884 0.804 Ave 0.819 0.777 0.692 0.763 0.711 0.824 0.788 0.845 0.777 The highest value and its method
  51. 51. C1 C2 C3 C4 C5 C6 V A Ave Comb V-Adj 0.781 Adj 0.869 All 0.713 N-Adj 0.780 N-Adj 0.750 V-Adj 0.856 All 0.783 Aj-Av 0.844 0.797 Music N-Adj 0.754 All 0.671 N-Adj 0.612 N-Adj 0.750 All2 0.725 All2 0.787 N-Adj 0.740 N-Adj 0.806 0.730 Video V-Adj 0.921 N-Adj 0.792 All 0.752 Adj 0.759 N-Adj 0.657 Adj 0.829 Aj-Av 0.840 Aj-Av 0.884 0.804 Ave 0.819 0.777 0.692 0.763 0.711 0.824 0.788 0.845 0.777 The highest value and its method Estimation from social comments is effective for C1, C6 and Arousal.
  52. 52. C1 C2 C3 C4 C5 C6 V A Ave Comb V-Adj 0.781 Adj 0.869 All 0.713 N-Adj 0.780 N-Adj 0.750 V-Adj 0.856 All 0.783 Aj-Av 0.844 0.797 Music N-Adj 0.754 All 0.671 N-Adj 0.612 N-Adj 0.750 All2 0.725 All2 0.787 N-Adj 0.740 N-Adj 0.806 0.730 Video V-Adj 0.921 N-Adj 0.792 All 0.752 Adj 0.759 N-Adj 0.657 Adj 0.829 Aj-Av 0.840 Aj-Av 0.884 0.804 Ave 0.819 0.777 0.692 0.763 0.711 0.824 0.788 0.845 0.777 The highest value and its method
  53. 53. Example of C6 (cute) MV C6 MV impression values = 2 It was able to learn well in SVM because the express impression words are similar.
  54. 54. C1 C2 C3 C4 C5 C6 V A Ave Comb V-Adj 0.781 Adj 0.869 All 0.713 N-Adj 0.780 N-Adj 0.750 V-Adj 0.856 All 0.783 Aj-Av 0.844 0.797 Music N-Adj 0.754 All 0.671 N-Adj 0.612 N-Adj 0.750 All2 0.725 All2 0.787 N-Adj 0.740 N-Adj 0.806 0.730 Video V-Adj 0.921 N-Adj 0.792 All 0.752 Adj 0.759 N-Adj 0.657 Adj 0.829 Aj-Av 0.840 Aj-Av 0.884 0.804 Ave 0.819 0.777 0.692 0.763 0.711 0.824 0.788 0.845 0.777 The highest value and its method
  55. 55. C1 C2 C3 C4 C5 C6 V A Ave Comb V-Adj 0.781 Adj 0.869 All 0.713 N-Adj 0.780 N-Adj 0.750 V-Adj 0.856 All 0.783 Aj-Av 0.844 0.797 Music N-Adj 0.754 All 0.671 N-Adj 0.612 N-Adj 0.750 All2 0.725 All2 0.787 N-Adj 0.740 N-Adj 0.806 0.730 Video V-Adj 0.921 N-Adj 0.792 All 0.752 Adj 0.759 N-Adj 0.657 Adj 0.829 Aj-Av 0.840 Aj-Av 0.884 0.804 Ave 0.819 0.777 0.692 0.763 0.711 0.824 0.788 0.845 0.777 The highest value and its method C3 and C5 are hard to estimate impressions from social comments.
  56. 56. Example of C3 (painful) MV C3 MV impression value = 1 C3 Comb All 0.713 Music N-Adj 0.612 Video All 0.752 Ave 0.692 There is a possibility that a part-of-speech other than 4 have features.
  57. 57. Example of C5 (humorous) MV C5 MV I couldn’t successfully learn because there are a wide range of ways to receive from humorous. impression value = 2 impression value = 1
  58. 58. C1 C2 C3 C4 C5 C6 V A Ave Comb V-Adj 0.781 Adj 0.869 All 0.713 N-Adj 0.780 N-Adj 0.750 V-Adj 0.856 All 0.783 Aj-Av 0.844 0.797 Music N-Adj 0.754 All 0.671 N-Adj 0.612 N-Adj 0.750 All2 0.725 All2 0.787 N-Adj 0.740 N-Adj 0.806 0.730 Video V-Adj 0.921 N-Adj 0.792 All 0.752 Adj 0.759 N-Adj 0.657 Adj 0.829 Aj-Av 0.840 Aj-Av 0.884 0.804 Ave 0.819 0.777 0.692 0.763 0.711 0.824 0.788 0.845 0.777 The highest value and its method
  59. 59. C1 C2 C3 C4 C5 C6 V A Ave Comb V-Adj 0.781 Adj 0.869 All 0.713 N-Adj 0.780 N-Adj 0.750 V-Adj 0.856 All 0.783 Aj-Av 0.844 0.797 Music N-Adj 0.754 All 0.671 N-Adj 0.612 N-Adj 0.750 All2 0.725 All2 0.787 N-Adj 0.740 N-Adj 0.806 0.730 Video V-Adj 0.921 N-Adj 0.792 All 0.752 Adj 0.759 N-Adj 0.657 Adj 0.829 Aj-Av 0.840 Aj-Av 0.884 0.804 Ave 0.819 0.777 0.692 0.763 0.711 0.824 0.788 0.845 0.777 The highest value and its method Impressions of video picture only can be estimated from social comments.
  60. 60. • We analyzed the impression estimation accuracy of music video clips from social comments. Summary We revealed that it is better to use proper parts-of-speech in social comments depending on each media/impression type. We generated the impression dataset of chorus part. Accuracy of method including adjective is high. C1, C6 and Arousal > C3 and C5 Video picture only > Music only

×