Analyzing the Impact of Music, Posters, & Trailer Tonality Using Machine Learning
summary
Claims movie box office class prediction with 89% accuracy using poster features alone and plans to add trailer tonality scored out of 10. Uses data from TMDB and about 3,800 videos, downloads trailers, extracts visuals (brightness, saturation) and audio (spectral centroid, RMS) and feeds metadata plus adjusted budget/revenue ratio into a CNN with one convolutional layer and three fully connected layers, trained with MSE loss, Adam, and early stopping. Reports class labels (1–5), notes outliers and a planned cloud application for creators.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 3:00Movie success complexity
- 6:37Data collection & preprocessing
- 10:32Visual & audio features
- 13:45Model architecture & training
- 15:57Model performance & metrics
- 19:05Live predictions
- 22:13Regional & cultural differences
- 26:13Audience responses to sound
- 28:03Analysis correlations & limits
- 30:15CNNs for sound data
- 32:05Advertising costs research
- 33:55Outliers & model accuracy
- 34:59Q&A
related talks
46mBeyond Bag of Words A Practitioner’s Guide to Advanced NLP
23mPrediction of music pairwise preferences from facial expressions
01h 23mBeyond Sentiment Emotion Mining with Python and machine learning
01h 29mShow, Don’t Tell How Data Is Changing Hollywood
01h 00mAI Powered Uplift Modeling and Stream Forecasting in Music Industry
33mWhen Stand-up Comedy Meets Machine Learning
27mBeing Smart with Art
31mClassifying Audio into Types using Python