Project Proposal

This is our project proposal for our CS 4641 Group Project, which aims to predict movie success based on past trends. This report is the first interpretation of our project, and it will be used as a guide as we move forward. This website will be continously updated as needed throughout the project, and will be finalized at the end of the semester upon project completion.

View on GitHub

Proposal Video

Introduction and Background

The movie industry is difficult to navigate, and predicting a film’s success can mean the difference between huge profits or significant losses. Traditionally, factors like "star power" and marketing campaigns were used but often proved unreliable. With platforms like YouTube, new data sources provide better insights into audience engagement. This project will use machine learning to predict movie success by combining traditional metadata with transmedial data.

Literature Review

Methods for predicting movie success previously exist; Early studies used sentiment analysis and data mining to forecast box office revenue based on factors like cast, budget, and genre [1]. More recent research examines how media sites impact movie popularity [2], and studies show YouTube trailer views can improve prediction accuracy [3]. However, combining traditional metadata with web data for newer films remains rather underexplored.

Dataset Description and Link

This project uses the “The Movies Dataset” from Kaggle [4], which contains movie metadata including genre, budget, revenue, and ratings. We will seek to incorporate features like YouTube trailer view counts and social media sentiment analysis, time permitting. We will concentrate on movies released up until July 2017 due to the recency of the dataset.

Access the dataset here.

Problem Definition

Studios and distributors require accurate predictions to guide decisions on marketing, distribution, and investment. Current methods often lack the precision needed in today’s evolving media landscape. This project will address this issue using machine learning techniques.

Motivation

Accurate movie predictions offer significant benefits. Studios can optimize budgets, distributors can refine release schedules, and investors can assess project viability. Additionally, understanding what drives success offers valuable insights into audience trends and preferences.

Methods

Data Preprocessing Methods

• Data Cleaning

  Handling missing values in the dataset using imputation techniques. (e.g., SciKit Imputation).

• Feature Engineering

  Potentially creating new features like “like-view ratio” and extract trailer view counts from YouTube.

• Data Transformation

  Scaling and normalizing numerical features using standardization and min-max scaling. (e.g., SciKit Transformers).

• One-Hot Encoding

  Transforming categorical variables into binary vectors for model processing.

Machine Learning Algorithms

• Random Forest Regression

  Random Forest can handle complex (non-linear and multiple-class) relationships between features like budget, cast, and social media engagement.

• Gradient Boosting Regression

  Gradient Boosting can enhance the model's accuracy by focusing on hard-to-predict movie success, making it especially effective for supposed outliers in the dataset.

• SVM (Support Vector Machine)

  Can classify/separate film data as hits or flops based on features aforementioned.

• kNN (K-Nearest Neighbors)

  By comparing new films to similar past releases, kNN can provide intuitive predictions with a distance function.

Learning Methods

This project primarily utilizes supervised learning techniques to train our models with labeled data from the dataset.

Analyzing Results and Discussion

Quantitative Metrics

• Coefficient of Determination (R^2)

  A strong R² would demonstrate how well factors like budget and cast explain a film’s financial success.

• Brier Score

  Beneficial for assessing the likelihood of a movie exceeding specific revenue thresholds, which could help studios/investors make more informed decisions.

• Class Likelihood Ratio

  By comparing the positive/negative likelihood ratios of binary classifiers, this metric would help the model make more confident predictions of movies as hits or flops.

• Log Loss

  By penalizing overconfident incorrect predictions, this metric helps ensure the model provides reliable probability estimates when classifying movies as hits or flops.

Project Goals

Primary goal is to develop a model that predicts movie success accurately. Ethically we will ensure data privacy, and sustainably we will aim to develop a model that can be updated with ease.

Expected Results

We expect to identify the most influential dataset features and expect that incorporating features like YouTube trailer views will significantly improve prediction accuracy.

References

1. Quader, N., et al. “A Machine Learning Approach to Predict Movie Box-Office Success.” IEEE Xplore, 1 Dec. 2017, ieeexplore.ieee.org/document/8281839 .

2. him, Steve, and Mohammad Pourhomayoun. Predicting Movie Market Revenue Using Social Media Data. 1 Aug. 2017, pp. 478–484, ieeexplore.ieee.org/abstract/document/8102973 , https://doi.org/10.1109/iri.2017.68 . Accessed 21 Feb. 2025.
3. Ahmad, Ibrahim Said, et al. “Movie Revenue Prediction Based on Purchase Intention Mining Using YouTube Trailer Reviews.” Information Processing & Management, vol. 57, no. 5, 1 Sept. 2020, p. 102278, www.sciencedirect.com/science/article/abs/pii/S0306457319309501 , https://doi.org/10.1016/j.ipm.2020.102278
4. Rounak Banik. The Movies Dataset. Kaggle. https://www.kaggle.com/datasets/rounakbanik/the-movies-dataset

---

Proposal Contributions

Dylan Bruce	Timotheus James	Vikrant Talwar	Alexander Thorne	Tyler Morgan
Established GitHub repository/website with Jekyll theme, researched datasets, identified data preprocessing methods & ML algorithms, created video slides, and translated proposal writeup to website.	Researched background, literature review, explicitly defined the problem definition, ensured IEEE format.	Identified problem motivation, data preprocessing methods, ML algorithms, and recorded/uploaded proposal video.	Edited Project Proposal, dataset research.	Identified strategies for preprocessing data, and explored model architectures.

---

Gantt Chart

Click to access project Gantt Chart