zdravinho/nwe

Nuwe_Challenge

➡️ Context: Welcome to the Lights, Camera, Action Challenge! 🌟🎥 In this thrilling competition, you will dive into the world of movies and TV shows, predicting the IMDB score classes like a seasoned film critic. The IMDB score is a well-established metric in the film industry, serving as a barometer for a title's quality. Ready for your close-up? Let's roll!

✍️ Dataset: For this challenge, you will have 4 csv files: train_titles.csv, train_credits.csv, test_titles.csv, and test_credits.csv.

In train_titles.csv and test_titles.csv, each row is a unique title, with the following scenes:

id: A unique identifier for each title. title: The name of the title. type: The type of the title (e.g., movie, show). description: The synopsis of the title. release_year: The year the title was released. age_certification: The age certification of the title. runtime: The runtime of the title. seasons: The number of seasons (relevant for shows). genres: The list of genres of the title. production_countries: The list of countries where the title was produced. streaming: The streaming platform offering the title. imdb_votes: The number of IMDB votes the title has received. The train_titles.csv file also includes:

imdb_score: The IMDB score of the title. target: The score class of the title. This is the transformation of the IMDB score using the following method: def get_imdb_score_class(score): if score < 2: # 1 <= score < 2 return 0 elif score < 3: return 1 elif score < 4: return 2 elif score < 5: return 3 elif score < 6: return 4 elif score < 7: return 5 elif score < 8: return 6 elif score < 9: return 7 else: # 9 <= score <= 10 return 8

Using pandas apply method

train_titles['target'] = train_titles['imdb_score'].apply(get_imdb_score_class) The train_credits.csv and test_credits.csv files are your supporting cast, providing extra information about the actors and directors of each title:

id: The identifier of the title, corresponding to the id in the titles datasets. person_id: A unique identifier for each person. name: The name of the person. character: The name of the character played in the film. role: The role of the person (e.g., actor, director). Although adding the titles credits information can boost your model performance, it is not required to use the credits datasets.

🎯 Objectives and tasks: Your only objective is:

Predicting the IMDB score for the test titles. Please note that your model should predict the IMDB score, and then transform this score into the target column using the get_imdb_score_class function provided.

✍️ Evaluation & submission format Your performance will be judged using the F1 macro score with the target column.