SlideShare ist ein Scribd-Unternehmen logo
1 von 23
Downloaden Sie, um offline zu lesen
Music Recommandation System
         Group - G6
  Advisor – Dr. Vikram Pudi
Music != Movies and books
•   CF algorithms are generally suffers from cold-
  start problem, novelty and ignore content of
  items.
• Tracking user’s preference is mostly done
  implicitly, via their listening habits instead
  of asking users to explicitly rate the item
• Any user can consume the item several times,
  even repeatedly and continuously. Mostly music
  labeling should be done by music experts.
• Another big difference is context of the music,
  like people prefer hard-rock in the morning,
  classical piano while working, and cool jazz
  while having dinner. It should also handle these
  contextual differences
Audio Feature Extraction

• Audio data is the time series where y-axis is
  current amplitude and x-axis is time.
Continue ..
• Audio waveform is broken into short frames.
  (1024 samples at 22050Hz).
• Collect Frame-level features and get
  mean/variance for each frames
• Discrete short term Fourier transformation
• Real Cepstral Coefficients
• Mel Frequency Cepstral Coefficients
• Zero crossing rate
• Septral centroid, Rolloff, flux and LPC
• Rhythmic and Harmony Beat features ..
• We got a 68 floating point vector called
  feature vector for a audio file.
• Computationally expensive
2. Automatic Playlist Generation using
              a song seed
• Aim:- given a song suggest most similar
  song in the library and make a mood-based
  playlist
• Implementation-
• Seed song s0 , F0 = {p1 , p2 , …………., pn}
• Find song s { S- s0} where difference
  between feature vector is minimum.
• Repeat the steps to generate whole playlist
  up to certain tolerance.
• Two mode:- seed song, last recommended song
• Note that there is no user involved.
Problem in User scenario

• Previous was simplest case of recommendation
  problem.
• No user preferences are involved.
• Where is user profiles , musical taste,
  feedback loops, ratings, listening habits ?
• Result depends upon seed song, all the time
• Not ideal situation when user has various
  list of songs already in his playlist.
• Simple Averaging can’t be the right solution
3. Top-N recommendation
• Solution: Clustering with dynamic K-mean
• Cluster the song using kmean algorithm based
  on their feature vectors.
• But we don’t know the initial K ? Solution
• Fix a Radius R at which a genre is usually
  clustered. If user like 3 genres, finally 3
  or more clusters will be the outcome
• Algorithm starts with k=1, if radius > R,
  increase K by 1 (=2) and recalculate until
  all cluster’s radius <= R.
• Then find score of each music and select
  top-N items, N is given by user.
Ranking and scoring items
•   Calculate score of each music as
•
•   Score(m,c) = 1 * ClusterData(c)
                  --------------------------
                   Dist( Vc , Vm ) * AllData

•   score(m,c) = score of music item with cluster c,
•   ClusterData(c) = number of music instances in cluster
    c,
•   Dist(v,u) = Euclidian distance between music item and
    cluster centroid. Hence more closer to the centroid,
    more chances of getting recommended. i.e. higher score.
•   Alldata = number of pieces in users’ playlist.
•   Sum the score for each cluster.
•   Sort down the score and recommend top-N items.
Stats
• Dataset = 'A benchmark for automatic genre
  classification"
• +----------+--------+
•   | tag         | count(*) |
•   +-------------+---------+
•   | alternative | 145      |
•   | blues       | 120      |
•   | electronic | 113       |
•   | folkcountry | 222      |
•   | funksoulrnb | 47       |
•   | jazz        | 319      |
•   | pop         | 116      |
•   | raphiphop   | 300      |
•   | rock        | 504      |
•   +-------------+----------+
Limitation
• No user feedback loops
• Determining R is a problem, this approach
  fails in case of numerous genres.
• Content based recommendation are less
  accurate .
• Hard –time of mapping user preferences into
  music domain.
• More feature will increase the result but
  clustering is a expensive with big feature
  vectors.
• Scaling problem, not efficient.
Future work
• UI front-end and Interfaces
• Improving recommendation algorithm defined in
  literature.
• Gathering implicit feedback and tracking user-
  profiles.
• Playlists according to artists, user-profile as a
  seed.
• Recommendation from a song-set.
• Mining web for new music information and other
  attributes of songs. Mp3 blogs, web services API
  last.fm, mystrands, pandora etc.
References
•   Adomavicius, G. and Tuzhilin A.(2005), “Towards the
    next generation of recommender system: A survey and
    state-of-art and possible extensions”
•   Aucouturier, J-J. and Packet, F. (2002), “Music
    similarity measures: What’s the use?”
•   P. Cano, M. Kopperbergerger, N. Wack, “Content based
    music audio recommendation”
•   B. Logan, “Music recommendation from song-sets”
•   G. Tzanetakis, P. Cook, “Musical genre classification
    of audio signals”
•   J.H. Ban, K.M. Kim, K.S. Park, “Quick audio retrieval
    using multiple feature vector”
•    Canno, P., Koppenberger, M., and Wack, N. (2005), “An
    industrial-strength content based music recommendation
    system   ”

Weitere ähnliche Inhalte

Andere mochten auch

SMP2014 - DEMINO - RUSSIAN INTRO
SMP2014 - DEMINO - RUSSIAN INTROSMP2014 - DEMINO - RUSSIAN INTRO
SMP2014 - DEMINO - RUSSIAN INTROKhryashchev
 
Номинации 2014
Номинации 2014Номинации 2014
Номинации 2014Khryashchev
 
2012Piclab_summary
2012Piclab_summary2012Piclab_summary
2012Piclab_summaryKhryashchev
 
Future of Digital TV
Future of Digital TVFuture of Digital TV
Future of Digital TVKhryashchev
 
Level Design Slides
Level Design SlidesLevel Design Slides
Level Design Slidesjscott4444
 
Eurocon2009 Apalkov
Eurocon2009 ApalkovEurocon2009 Apalkov
Eurocon2009 ApalkovKhryashchev
 
A Short PMML Tutorial by LatentView
A Short PMML Tutorial by LatentViewA Short PMML Tutorial by LatentView
A Short PMML Tutorial by LatentViewramesh.latentview
 
Poverty in India
Poverty in IndiaPoverty in India
Poverty in IndiaAnirban DG
 

Andere mochten auch (11)

Portfolio
PortfolioPortfolio
Portfolio
 
SMP2014 - DEMINO - RUSSIAN INTRO
SMP2014 - DEMINO - RUSSIAN INTROSMP2014 - DEMINO - RUSSIAN INTRO
SMP2014 - DEMINO - RUSSIAN INTRO
 
1
11
1
 
Номинации 2014
Номинации 2014Номинации 2014
Номинации 2014
 
2012Piclab_summary
2012Piclab_summary2012Piclab_summary
2012Piclab_summary
 
Future of Digital TV
Future of Digital TVFuture of Digital TV
Future of Digital TV
 
Level Design Slides
Level Design SlidesLevel Design Slides
Level Design Slides
 
Kuykin_diploma
Kuykin_diplomaKuykin_diploma
Kuykin_diploma
 
Eurocon2009 Apalkov
Eurocon2009 ApalkovEurocon2009 Apalkov
Eurocon2009 Apalkov
 
A Short PMML Tutorial by LatentView
A Short PMML Tutorial by LatentViewA Short PMML Tutorial by LatentView
A Short PMML Tutorial by LatentView
 
Poverty in India
Poverty in IndiaPoverty in India
Poverty in India
 

Ähnlich wie Btp 1st

Personalized Music Emotion Recognition via Model Adaptation
Personalized Music Emotion Recognition via Model AdaptationPersonalized Music Emotion Recognition via Model Adaptation
Personalized Music Emotion Recognition via Model AdaptationJu-Chiang Wang
 
The Acoustic Emotion Gaussians Model for Emotion-based Music Annotation and R...
The Acoustic Emotion Gaussians Model for Emotion-based Music Annotation and R...The Acoustic Emotion Gaussians Model for Emotion-based Music Annotation and R...
The Acoustic Emotion Gaussians Model for Emotion-based Music Annotation and R...Ju-Chiang Wang
 
The Secret Lives of MP3 Files
The Secret Lives of MP3 FilesThe Secret Lives of MP3 Files
The Secret Lives of MP3 Filesdkaye
 
Music Recommendation and Discovery in...which Web?
Music Recommendation and Discovery in...which Web?Music Recommendation and Discovery in...which Web?
Music Recommendation and Discovery in...which Web?Oscar Celma
 
The Caspar Artistic Testbed Jerome Barthelemy
The Caspar Artistic Testbed Jerome BarthelemyThe Caspar Artistic Testbed Jerome Barthelemy
The Caspar Artistic Testbed Jerome BarthelemyDigitalPreservationEurope
 
Music similarity: what for?
Music similarity: what for?Music similarity: what for?
Music similarity: what for?Emilia Gómez
 
Automatic Set List Identification and Song Segmentation of Full-Length Concer...
Automatic Set List Identification and Song Segmentation of Full-Length Concer...Automatic Set List Identification and Song Segmentation of Full-Length Concer...
Automatic Set List Identification and Song Segmentation of Full-Length Concer...Ju-Chiang Wang
 
Deep Learning Meetup #5
Deep Learning Meetup #5Deep Learning Meetup #5
Deep Learning Meetup #5Aloïs Gruson
 
Research at MAC Lab, Academia Sincia, in 2017
Research at MAC Lab, Academia Sincia, in 2017Research at MAC Lab, Academia Sincia, in 2017
Research at MAC Lab, Academia Sincia, in 2017Yi-Hsuan Yang
 
Audio Separation Comparison: Clustering Repeating Period and Hidden Markov Model
Audio Separation Comparison: Clustering Repeating Period and Hidden Markov ModelAudio Separation Comparison: Clustering Repeating Period and Hidden Markov Model
Audio Separation Comparison: Clustering Repeating Period and Hidden Markov ModelYao Yao
 
Timbral modeling for music artist recognition using i-vectors
Timbral modeling for music artist recognition using i-vectorsTimbral modeling for music artist recognition using i-vectors
Timbral modeling for music artist recognition using i-vectorsHamid Eghbal-zadeh
 
Online Recommender System for Radio Station Hosting: Experimental Results Rev...
Online Recommender System for Radio Station Hosting: Experimental Results Rev...Online Recommender System for Radio Station Hosting: Experimental Results Rev...
Online Recommender System for Radio Station Hosting: Experimental Results Rev...Dmitrii Ignatov
 
[221]똑똑한 인공지능 dj 비서 clova music
[221]똑똑한 인공지능 dj 비서 clova music[221]똑똑한 인공지능 dj 비서 clova music
[221]똑똑한 인공지능 dj 비서 clova musicNAVER D2
 
Music discovery on the net
Music discovery on the netMusic discovery on the net
Music discovery on the netguestbf080
 
Recsys Challenge 2018 - Creamy Fireflies - Artist-driven layering and user’s...
Recsys Challenge 2018 - Creamy Fireflies -  Artist-driven layering and user’s...Recsys Challenge 2018 - Creamy Fireflies -  Artist-driven layering and user’s...
Recsys Challenge 2018 - Creamy Fireflies - Artist-driven layering and user’s...Emanuele Chioso
 
Random Forests R vs Python by Linda Uruchurtu
Random Forests R vs Python by Linda UruchurtuRandom Forests R vs Python by Linda Uruchurtu
Random Forests R vs Python by Linda UruchurtuPyData
 
ScoReader: A Mobile Computer Vision System for Optical Music Recognition
ScoReader: A Mobile Computer Vision System for Optical Music RecognitionScoReader: A Mobile Computer Vision System for Optical Music Recognition
ScoReader: A Mobile Computer Vision System for Optical Music Recognitionchaser55
 
Mining the social web for music-related data: a hands-on tutorial
Mining the social web for music-related data: a hands-on tutorialMining the social web for music-related data: a hands-on tutorial
Mining the social web for music-related data: a hands-on tutorialBen Fields
 

Ähnlich wie Btp 1st (20)

Personalized Music Emotion Recognition via Model Adaptation
Personalized Music Emotion Recognition via Model AdaptationPersonalized Music Emotion Recognition via Model Adaptation
Personalized Music Emotion Recognition via Model Adaptation
 
The Acoustic Emotion Gaussians Model for Emotion-based Music Annotation and R...
The Acoustic Emotion Gaussians Model for Emotion-based Music Annotation and R...The Acoustic Emotion Gaussians Model for Emotion-based Music Annotation and R...
The Acoustic Emotion Gaussians Model for Emotion-based Music Annotation and R...
 
The Secret Lives of MP3 Files
The Secret Lives of MP3 FilesThe Secret Lives of MP3 Files
The Secret Lives of MP3 Files
 
Music Recommendation and Discovery in...which Web?
Music Recommendation and Discovery in...which Web?Music Recommendation and Discovery in...which Web?
Music Recommendation and Discovery in...which Web?
 
The Caspar Artistic Testbed Jerome Barthelemy
The Caspar Artistic Testbed Jerome BarthelemyThe Caspar Artistic Testbed Jerome Barthelemy
The Caspar Artistic Testbed Jerome Barthelemy
 
Music similarity: what for?
Music similarity: what for?Music similarity: what for?
Music similarity: what for?
 
Automatic Set List Identification and Song Segmentation of Full-Length Concer...
Automatic Set List Identification and Song Segmentation of Full-Length Concer...Automatic Set List Identification and Song Segmentation of Full-Length Concer...
Automatic Set List Identification and Song Segmentation of Full-Length Concer...
 
Deep Learning Meetup #5
Deep Learning Meetup #5Deep Learning Meetup #5
Deep Learning Meetup #5
 
Research at MAC Lab, Academia Sincia, in 2017
Research at MAC Lab, Academia Sincia, in 2017Research at MAC Lab, Academia Sincia, in 2017
Research at MAC Lab, Academia Sincia, in 2017
 
Audio Separation Comparison: Clustering Repeating Period and Hidden Markov Model
Audio Separation Comparison: Clustering Repeating Period and Hidden Markov ModelAudio Separation Comparison: Clustering Repeating Period and Hidden Markov Model
Audio Separation Comparison: Clustering Repeating Period and Hidden Markov Model
 
Timbral modeling for music artist recognition using i-vectors
Timbral modeling for music artist recognition using i-vectorsTimbral modeling for music artist recognition using i-vectors
Timbral modeling for music artist recognition using i-vectors
 
Online Recommender System for Radio Station Hosting: Experimental Results Rev...
Online Recommender System for Radio Station Hosting: Experimental Results Rev...Online Recommender System for Radio Station Hosting: Experimental Results Rev...
Online Recommender System for Radio Station Hosting: Experimental Results Rev...
 
[221]똑똑한 인공지능 dj 비서 clova music
[221]똑똑한 인공지능 dj 비서 clova music[221]똑똑한 인공지능 dj 비서 clova music
[221]똑똑한 인공지능 dj 비서 clova music
 
Music discovery on the net
Music discovery on the netMusic discovery on the net
Music discovery on the net
 
We love NLTK
We love NLTKWe love NLTK
We love NLTK
 
Recsys Challenge 2018 - Creamy Fireflies - Artist-driven layering and user’s...
Recsys Challenge 2018 - Creamy Fireflies -  Artist-driven layering and user’s...Recsys Challenge 2018 - Creamy Fireflies -  Artist-driven layering and user’s...
Recsys Challenge 2018 - Creamy Fireflies - Artist-driven layering and user’s...
 
10-02-ACA-Alignment.pdf
10-02-ACA-Alignment.pdf10-02-ACA-Alignment.pdf
10-02-ACA-Alignment.pdf
 
Random Forests R vs Python by Linda Uruchurtu
Random Forests R vs Python by Linda UruchurtuRandom Forests R vs Python by Linda Uruchurtu
Random Forests R vs Python by Linda Uruchurtu
 
ScoReader: A Mobile Computer Vision System for Optical Music Recognition
ScoReader: A Mobile Computer Vision System for Optical Music RecognitionScoReader: A Mobile Computer Vision System for Optical Music Recognition
ScoReader: A Mobile Computer Vision System for Optical Music Recognition
 
Mining the social web for music-related data: a hands-on tutorial
Mining the social web for music-related data: a hands-on tutorialMining the social web for music-related data: a hands-on tutorial
Mining the social web for music-related data: a hands-on tutorial
 

Kürzlich hochgeladen

Genesis 1:5 - Meditate the Scripture Daily bit by bit
Genesis 1:5 - Meditate the Scripture Daily bit by bitGenesis 1:5 - Meditate the Scripture Daily bit by bit
Genesis 1:5 - Meditate the Scripture Daily bit by bitmaricelcanoynuay
 
Genesis 1:10 || Meditate the Scripture daily verse by verse
Genesis 1:10  ||  Meditate the Scripture daily verse by verseGenesis 1:10  ||  Meditate the Scripture daily verse by verse
Genesis 1:10 || Meditate the Scripture daily verse by versemaricelcanoynuay
 
Deerfoot Church of Christ Bulletin 5 5 24
Deerfoot Church of Christ Bulletin 5 5 24Deerfoot Church of Christ Bulletin 5 5 24
Deerfoot Church of Christ Bulletin 5 5 24deerfootcoc
 
St John's Church Parish Diary for May 2024
St John's Church Parish Diary for May 2024St John's Church Parish Diary for May 2024
St John's Church Parish Diary for May 2024Chris Lyne
 
Emails, Facebook, WhatsApp and the Dhamma (English and Chinese).pdf
Emails, Facebook, WhatsApp and the Dhamma  (English and Chinese).pdfEmails, Facebook, WhatsApp and the Dhamma  (English and Chinese).pdf
Emails, Facebook, WhatsApp and the Dhamma (English and Chinese).pdfOH TEIK BIN
 
Amil baba in Lahore /Amil baba in Karachi /Amil baba in Pakistan
Amil baba in Lahore /Amil baba in Karachi /Amil baba in PakistanAmil baba in Lahore /Amil baba in Karachi /Amil baba in Pakistan
Amil baba in Lahore /Amil baba in Karachi /Amil baba in PakistanAmil Baba Mangal Maseeh
 
Best Astrologer Vashikaran Specialist in Germany and France Black Magic Exper...
Best Astrologer Vashikaran Specialist in Germany and France Black Magic Exper...Best Astrologer Vashikaran Specialist in Germany and France Black Magic Exper...
Best Astrologer Vashikaran Specialist in Germany and France Black Magic Exper...Amil Baba Naveed Bangali
 
The_Chronological_Life_of_Christ_Part_99_Words_and_Works
The_Chronological_Life_of_Christ_Part_99_Words_and_WorksThe_Chronological_Life_of_Christ_Part_99_Words_and_Works
The_Chronological_Life_of_Christ_Part_99_Words_and_WorksNetwork Bible Fellowship
 
"The Magnificent Surah Rahman: PDF Version"
"The Magnificent Surah Rahman: PDF Version""The Magnificent Surah Rahman: PDF Version"
"The Magnificent Surah Rahman: PDF Version"aijazuddin14
 
Authentic Black magic, Kala ilam expert in UAE and Kala ilam specialist in S...
Authentic Black magic, Kala ilam expert in UAE  and Kala ilam specialist in S...Authentic Black magic, Kala ilam expert in UAE  and Kala ilam specialist in S...
Authentic Black magic, Kala ilam expert in UAE and Kala ilam specialist in S...baharayali
 
Connaught Place, Delhi Call girls :8448380779 Model Escorts | 100% verified
Connaught Place, Delhi Call girls :8448380779 Model Escorts | 100% verifiedConnaught Place, Delhi Call girls :8448380779 Model Escorts | 100% verified
Connaught Place, Delhi Call girls :8448380779 Model Escorts | 100% verifiedDelhi Call girls
 
Codex Singularity: Search for the Prisca Sapientia
Codex Singularity: Search for the Prisca SapientiaCodex Singularity: Search for the Prisca Sapientia
Codex Singularity: Search for the Prisca Sapientiajfrenchau
 
Top 10 Amil baba list Famous Amil baba In Pakistan Amil baba Kala jadu in Raw...
Top 10 Amil baba list Famous Amil baba In Pakistan Amil baba Kala jadu in Raw...Top 10 Amil baba list Famous Amil baba In Pakistan Amil baba Kala jadu in Raw...
Top 10 Amil baba list Famous Amil baba In Pakistan Amil baba Kala jadu in Raw...Amil Baba Naveed Bangali
 
MEIDUNIDADE COM JESUS PALESTRA ESPIRITA1.pptx
MEIDUNIDADE COM JESUS  PALESTRA ESPIRITA1.pptxMEIDUNIDADE COM JESUS  PALESTRA ESPIRITA1.pptx
MEIDUNIDADE COM JESUS PALESTRA ESPIRITA1.pptxMneasEntidades
 
Genesis 1:2 - Meditate the Scripture Daily bit by bit
Genesis 1:2 - Meditate the Scripture Daily bit by bitGenesis 1:2 - Meditate the Scripture Daily bit by bit
Genesis 1:2 - Meditate the Scripture Daily bit by bitmaricelcanoynuay
 
Christian Charism Ministry - Manifestation of spiritual gifts within the chur...
Christian Charism Ministry - Manifestation of spiritual gifts within the chur...Christian Charism Ministry - Manifestation of spiritual gifts within the chur...
Christian Charism Ministry - Manifestation of spiritual gifts within the chur...mustardseed108
 
Popular Kala Jadu, Black magic specialist in Sialkot and Kala ilam specialist...
Popular Kala Jadu, Black magic specialist in Sialkot and Kala ilam specialist...Popular Kala Jadu, Black magic specialist in Sialkot and Kala ilam specialist...
Popular Kala Jadu, Black magic specialist in Sialkot and Kala ilam specialist...baharayali
 

Kürzlich hochgeladen (20)

Famous No -1 amil baba in Hyderabad ! Best No _ Astrologer in Pakistan, UK, A...
Famous No -1 amil baba in Hyderabad ! Best No _ Astrologer in Pakistan, UK, A...Famous No -1 amil baba in Hyderabad ! Best No _ Astrologer in Pakistan, UK, A...
Famous No -1 amil baba in Hyderabad ! Best No _ Astrologer in Pakistan, UK, A...
 
Genesis 1:5 - Meditate the Scripture Daily bit by bit
Genesis 1:5 - Meditate the Scripture Daily bit by bitGenesis 1:5 - Meditate the Scripture Daily bit by bit
Genesis 1:5 - Meditate the Scripture Daily bit by bit
 
Genesis 1:10 || Meditate the Scripture daily verse by verse
Genesis 1:10  ||  Meditate the Scripture daily verse by verseGenesis 1:10  ||  Meditate the Scripture daily verse by verse
Genesis 1:10 || Meditate the Scripture daily verse by verse
 
Deerfoot Church of Christ Bulletin 5 5 24
Deerfoot Church of Christ Bulletin 5 5 24Deerfoot Church of Christ Bulletin 5 5 24
Deerfoot Church of Christ Bulletin 5 5 24
 
St John's Church Parish Diary for May 2024
St John's Church Parish Diary for May 2024St John's Church Parish Diary for May 2024
St John's Church Parish Diary for May 2024
 
Emails, Facebook, WhatsApp and the Dhamma (English and Chinese).pdf
Emails, Facebook, WhatsApp and the Dhamma  (English and Chinese).pdfEmails, Facebook, WhatsApp and the Dhamma  (English and Chinese).pdf
Emails, Facebook, WhatsApp and the Dhamma (English and Chinese).pdf
 
Amil baba in Lahore /Amil baba in Karachi /Amil baba in Pakistan
Amil baba in Lahore /Amil baba in Karachi /Amil baba in PakistanAmil baba in Lahore /Amil baba in Karachi /Amil baba in Pakistan
Amil baba in Lahore /Amil baba in Karachi /Amil baba in Pakistan
 
Best Astrologer Vashikaran Specialist in Germany and France Black Magic Exper...
Best Astrologer Vashikaran Specialist in Germany and France Black Magic Exper...Best Astrologer Vashikaran Specialist in Germany and France Black Magic Exper...
Best Astrologer Vashikaran Specialist in Germany and France Black Magic Exper...
 
The_Chronological_Life_of_Christ_Part_99_Words_and_Works
The_Chronological_Life_of_Christ_Part_99_Words_and_WorksThe_Chronological_Life_of_Christ_Part_99_Words_and_Works
The_Chronological_Life_of_Christ_Part_99_Words_and_Works
 
"The Magnificent Surah Rahman: PDF Version"
"The Magnificent Surah Rahman: PDF Version""The Magnificent Surah Rahman: PDF Version"
"The Magnificent Surah Rahman: PDF Version"
 
Authentic Black magic, Kala ilam expert in UAE and Kala ilam specialist in S...
Authentic Black magic, Kala ilam expert in UAE  and Kala ilam specialist in S...Authentic Black magic, Kala ilam expert in UAE  and Kala ilam specialist in S...
Authentic Black magic, Kala ilam expert in UAE and Kala ilam specialist in S...
 
Connaught Place, Delhi Call girls :8448380779 Model Escorts | 100% verified
Connaught Place, Delhi Call girls :8448380779 Model Escorts | 100% verifiedConnaught Place, Delhi Call girls :8448380779 Model Escorts | 100% verified
Connaught Place, Delhi Call girls :8448380779 Model Escorts | 100% verified
 
Codex Singularity: Search for the Prisca Sapientia
Codex Singularity: Search for the Prisca SapientiaCodex Singularity: Search for the Prisca Sapientia
Codex Singularity: Search for the Prisca Sapientia
 
Top 10 Amil baba list Famous Amil baba In Pakistan Amil baba Kala jadu in Raw...
Top 10 Amil baba list Famous Amil baba In Pakistan Amil baba Kala jadu in Raw...Top 10 Amil baba list Famous Amil baba In Pakistan Amil baba Kala jadu in Raw...
Top 10 Amil baba list Famous Amil baba In Pakistan Amil baba Kala jadu in Raw...
 
St. Louise de Marillac and Poor Children
St. Louise de Marillac and Poor ChildrenSt. Louise de Marillac and Poor Children
St. Louise de Marillac and Poor Children
 
St. Louise de Marillac and Abandoned Children
St. Louise de Marillac and Abandoned ChildrenSt. Louise de Marillac and Abandoned Children
St. Louise de Marillac and Abandoned Children
 
MEIDUNIDADE COM JESUS PALESTRA ESPIRITA1.pptx
MEIDUNIDADE COM JESUS  PALESTRA ESPIRITA1.pptxMEIDUNIDADE COM JESUS  PALESTRA ESPIRITA1.pptx
MEIDUNIDADE COM JESUS PALESTRA ESPIRITA1.pptx
 
Genesis 1:2 - Meditate the Scripture Daily bit by bit
Genesis 1:2 - Meditate the Scripture Daily bit by bitGenesis 1:2 - Meditate the Scripture Daily bit by bit
Genesis 1:2 - Meditate the Scripture Daily bit by bit
 
Christian Charism Ministry - Manifestation of spiritual gifts within the chur...
Christian Charism Ministry - Manifestation of spiritual gifts within the chur...Christian Charism Ministry - Manifestation of spiritual gifts within the chur...
Christian Charism Ministry - Manifestation of spiritual gifts within the chur...
 
Popular Kala Jadu, Black magic specialist in Sialkot and Kala ilam specialist...
Popular Kala Jadu, Black magic specialist in Sialkot and Kala ilam specialist...Popular Kala Jadu, Black magic specialist in Sialkot and Kala ilam specialist...
Popular Kala Jadu, Black magic specialist in Sialkot and Kala ilam specialist...
 

Btp 1st

  • 1. Music Recommandation System Group - G6 Advisor – Dr. Vikram Pudi
  • 2.
  • 3.
  • 4.
  • 5.
  • 6.
  • 7.
  • 8.
  • 9.
  • 10. Music != Movies and books • CF algorithms are generally suffers from cold- start problem, novelty and ignore content of items. • Tracking user’s preference is mostly done implicitly, via their listening habits instead of asking users to explicitly rate the item • Any user can consume the item several times, even repeatedly and continuously. Mostly music labeling should be done by music experts. • Another big difference is context of the music, like people prefer hard-rock in the morning, classical piano while working, and cool jazz while having dinner. It should also handle these contextual differences
  • 11.
  • 12.
  • 13.
  • 14. Audio Feature Extraction • Audio data is the time series where y-axis is current amplitude and x-axis is time.
  • 15. Continue .. • Audio waveform is broken into short frames. (1024 samples at 22050Hz). • Collect Frame-level features and get mean/variance for each frames • Discrete short term Fourier transformation • Real Cepstral Coefficients • Mel Frequency Cepstral Coefficients • Zero crossing rate • Septral centroid, Rolloff, flux and LPC • Rhythmic and Harmony Beat features .. • We got a 68 floating point vector called feature vector for a audio file. • Computationally expensive
  • 16. 2. Automatic Playlist Generation using a song seed • Aim:- given a song suggest most similar song in the library and make a mood-based playlist • Implementation- • Seed song s0 , F0 = {p1 , p2 , …………., pn} • Find song s { S- s0} where difference between feature vector is minimum. • Repeat the steps to generate whole playlist up to certain tolerance. • Two mode:- seed song, last recommended song • Note that there is no user involved.
  • 17. Problem in User scenario • Previous was simplest case of recommendation problem. • No user preferences are involved. • Where is user profiles , musical taste, feedback loops, ratings, listening habits ? • Result depends upon seed song, all the time • Not ideal situation when user has various list of songs already in his playlist. • Simple Averaging can’t be the right solution
  • 18. 3. Top-N recommendation • Solution: Clustering with dynamic K-mean • Cluster the song using kmean algorithm based on their feature vectors. • But we don’t know the initial K ? Solution • Fix a Radius R at which a genre is usually clustered. If user like 3 genres, finally 3 or more clusters will be the outcome • Algorithm starts with k=1, if radius > R, increase K by 1 (=2) and recalculate until all cluster’s radius <= R. • Then find score of each music and select top-N items, N is given by user.
  • 19. Ranking and scoring items • Calculate score of each music as • • Score(m,c) = 1 * ClusterData(c) -------------------------- Dist( Vc , Vm ) * AllData • score(m,c) = score of music item with cluster c, • ClusterData(c) = number of music instances in cluster c, • Dist(v,u) = Euclidian distance between music item and cluster centroid. Hence more closer to the centroid, more chances of getting recommended. i.e. higher score. • Alldata = number of pieces in users’ playlist. • Sum the score for each cluster. • Sort down the score and recommend top-N items.
  • 20. Stats • Dataset = 'A benchmark for automatic genre classification" • +----------+--------+ • | tag | count(*) | • +-------------+---------+ • | alternative | 145 | • | blues | 120 | • | electronic | 113 | • | folkcountry | 222 | • | funksoulrnb | 47 | • | jazz | 319 | • | pop | 116 | • | raphiphop | 300 | • | rock | 504 | • +-------------+----------+
  • 21. Limitation • No user feedback loops • Determining R is a problem, this approach fails in case of numerous genres. • Content based recommendation are less accurate . • Hard –time of mapping user preferences into music domain. • More feature will increase the result but clustering is a expensive with big feature vectors. • Scaling problem, not efficient.
  • 22. Future work • UI front-end and Interfaces • Improving recommendation algorithm defined in literature. • Gathering implicit feedback and tracking user- profiles. • Playlists according to artists, user-profile as a seed. • Recommendation from a song-set. • Mining web for new music information and other attributes of songs. Mp3 blogs, web services API last.fm, mystrands, pandora etc.
  • 23. References • Adomavicius, G. and Tuzhilin A.(2005), “Towards the next generation of recommender system: A survey and state-of-art and possible extensions” • Aucouturier, J-J. and Packet, F. (2002), “Music similarity measures: What’s the use?” • P. Cano, M. Kopperbergerger, N. Wack, “Content based music audio recommendation” • B. Logan, “Music recommendation from song-sets” • G. Tzanetakis, P. Cook, “Musical genre classification of audio signals” • J.H. Ban, K.M. Kim, K.S. Park, “Quick audio retrieval using multiple feature vector” • Canno, P., Koppenberger, M., and Wack, N. (2005), “An industrial-strength content based music recommendation system ”