{ "id": 245486, "title": "Compilation of Time Series Forecasting Competitions and Top Kernels", "url": "/competitions/mlb-player-digital-engagement-forecasting/discussion/245486", "author_name": "Athar Sayed", "post_date": "2021-06-11T05:44:25.875000", "votes": 8, "comment_count": 0, "views": 0, "content": "

Note: This competition involves prediction of Fans Engagement for MLB players’ digital content for future date range , this problem is like time series forecasting problem , hence ideas from previous kaggle competitions involving time series will be useful.

\n

Note: If you are new to time series to get started you may want to read articles mentioned https://www.kaggle.com/discussion/245201

\n

Here is the list of past time series competitions and along with that some top kernels

\n

1. Web Traffic Time Series Forecasting: Here task is to predict web traffic on web pages , note that this competition also involved evaluation of models on real time data , EDA Kernel here , Modelling kernel can be found here

\n

2. M5 Forecasting Accuracy: Aim is to predict unit sales of various products accross various stores in different locations , this competition involved hidden test over which no public LB feedback was given in last month . Hence ideas from it may be useful to build reliable validation strategy and robust models . Basic Lightgbm training kernel can be found here , Here is another very nice kernel for feature engineering and modelling

\n

3.Rossman Sales Store Competition:Classical Time series competition involving sales prediction , this was the competition which introduced categorical embeddings .

\n

This is are some of the competitions which can be useful
\nThanks for Reading !
\nHappy Kaggling.

", "messages": [ { "id": 1344757, "postDate": "2021-06-11T05:44:25.877Z", "content": "

Note: This competition involves prediction of Fans Engagement for MLB players’ digital content for future date range , this problem is like time series forecasting problem , hence ideas from previous kaggle competitions involving time series will be useful.

\n

Note: If you are new to time series to get started you may want to read articles mentioned https://www.kaggle.com/discussion/245201

\n

Here is the list of past time series competitions and along with that some top kernels

\n

1. Web Traffic Time Series Forecasting: Here task is to predict web traffic on web pages , note that this competition also involved evaluation of models on real time data , EDA Kernel here , Modelling kernel can be found here

\n

2. M5 Forecasting Accuracy: Aim is to predict unit sales of various products accross various stores in different locations , this competition involved hidden test over which no public LB feedback was given in last month . Hence ideas from it may be useful to build reliable validation strategy and robust models . Basic Lightgbm training kernel can be found here , Here is another very nice kernel for feature engineering and modelling

\n

3.Rossman Sales Store Competition:Classical Time series competition involving sales prediction , this was the competition which introduced categorical embeddings .

\n

This is are some of the competitions which can be useful
\nThanks for Reading !
\nHappy Kaggling.

", "rawMarkdown": "**Note: This competition involves prediction of Fans Engagement for MLB players’ digital content for future date range , this problem is like time series forecasting problem , hence ideas from previous kaggle competitions involving time series will be useful.**\n\nNote: If you are new to time series to get started you may want to read articles mentioned [https://www.kaggle.com/discussion/245201](url)\n\nHere is the list of past time series competitions and along with that some top kernels\n\n**[1. Web Traffic Time Series Forecasting](https://www.kaggle.com/c/web-traffic-time-series-forecasting/):** Here task is to predict web traffic on web pages , note that this competition also involved evaluation of models on real time data , EDA Kernel [here ](https://www.kaggle.com/muonneutrino/wikipedia-traffic-data-exploration) , Modelling kernel can be found [here](https://www.kaggle.com/zoupet/predictive-analysis-with-different-approaches)\n\n\n**[2. M5 Forecasting Accuracy](https://www.kaggle.com/c/m5-forecasting-accuracy):** Aim is to predict unit sales of various products accross various stores in different locations , this competition involved hidden test over which no public LB feedback was given in last month . Hence ideas from it may be useful to build reliable validation strategy and robust models . Basic Lightgbm training kernel can be found [here](https://www.kaggle.com/kyakovlev/m5-three-shades-of-dark-darker-magic) , [Here ](https://www.kaggle.com/kneroma/m5-first-public-notebook-under-0-50)is another very nice kernel for feature engineering and modelling\n\n**[3.Rossman Sales Store Competition](https://www.kaggle.com/c/rossmann-store-sales/overview):**Classical Time series competition involving sales prediction , this was the competition which introduced categorical embeddings .\n\nThis is are some of the competitions which can be useful\nThanks for Reading !\nHappy Kaggling.", "votes": 8 } ], "comments": [], "raw_markdown_by_id": { "1344757": "**Note: This competition involves prediction of Fans Engagement for MLB players’ digital content for future date range , this problem is like time series forecasting problem , hence ideas from previous kaggle competitions involving time series will be useful.**\n\nNote: If you are new to time series to get started you may want to read articles mentioned [https://www.kaggle.com/discussion/245201](url)\n\nHere is the list of past time series competitions and along with that some top kernels\n\n**[1. Web Traffic Time Series Forecasting](https://www.kaggle.com/c/web-traffic-time-series-forecasting/):** Here task is to predict web traffic on web pages , note that this competition also involved evaluation of models on real time data , EDA Kernel [here ](https://www.kaggle.com/muonneutrino/wikipedia-traffic-data-exploration) , Modelling kernel can be found [here](https://www.kaggle.com/zoupet/predictive-analysis-with-different-approaches)\n\n\n**[2. M5 Forecasting Accuracy](https://www.kaggle.com/c/m5-forecasting-accuracy):** Aim is to predict unit sales of various products accross various stores in different locations , this competition involved hidden test over which no public LB feedback was given in last month . Hence ideas from it may be useful to build reliable validation strategy and robust models . Basic Lightgbm training kernel can be found [here](https://www.kaggle.com/kyakovlev/m5-three-shades-of-dark-darker-magic) , [Here ](https://www.kaggle.com/kneroma/m5-first-public-notebook-under-0-50)is another very nice kernel for feature engineering and modelling\n\n**[3.Rossman Sales Store Competition](https://www.kaggle.com/c/rossmann-store-sales/overview):**Classical Time series competition involving sales prediction , this was the competition which introduced categorical embeddings .\n\nThis is are some of the competitions which can be useful\nThanks for Reading !\nHappy Kaggling." } }