ML-прогнозатор рекламы: как оценить результат до запуска — Давидько Владимир, Авито

Vladimir Davidko, in his presentation at Data Fest 2026, discussed how they built an ML advertising forecaster and the challenges and difficulties they encountered. 00:00:05 Introduction 00:00:50 Presentation Structure 00:01:27 Advertiser Account 00:03:15 Account Structure 00:04:41 First Solution 00:06:21 Problems with the First Solution 00:08:15 New Solution Based on CatBoost 00:11:03 Metrics and Challenges 00:12:22 Impact of Factors on Forecasts 00:13:00 Hourly Forecasting 00:13:58 Model Training 00:15:01 Testing and Results 00:17:06 Technical Improvements 00:18:25 Q&A AvitoTech is a team of Avito engineers. Subscribe to our channel, social media, and blogs to learn more about Avito technologies 👇🏻 VK: https://vk.com/avitotech Telegram: https://t.me/+aB2Kl5oJw-g5MzJi Habr: https://habr.com/ru/company/avito GitHub: https://github.com/avito-tech RuTube: https://rutube.ru/channel/30462632/ Zen: https://dzen.ru/avitotech Website: https://clc.to/l5ZqRA