Hey community i was going through this competition and people's notes and i am unable to understand few things
\n-> Isn't this a time series dataset , if yes i am seeing people using Lightgbm classifier ? , i read code and discussion topics of this competition and just need a few starting pointers like how to approach this dataset what models will be best to use here
Hey community i was going through this competition and people's notes and i am unable to understand few things
\n-> Isn't this a time series dataset , if yes i am seeing people using Lightgbm classifier ? , i read code and discussion topics of this competition and just need a few starting pointers like how to approach this dataset what models will be best to use here
Hey @g9jiggy, it's because the test dataset is shuffled. Only the training dataset provides timestamps, sorted by game_num and event_id.
", "rawMarkdown": "Hey @g9jiggy, it's because the test dataset is shuffled. Only the training dataset provides timestamps, sorted by game_num and event_id.", "votes": null }, { "id": "1968015", "postDate": "10/02/2022 20:05:02", "content": "You may use such models for this dataset, this dataset contains temporal information along with much more. Also @aatiffraz has correctly pointed out that the test set is randomly shuffled leading to a loss of time-based continuity.
\nHope this helps @g9jiggy
Thanks @aatiffraz and @ravi20076 , so this is a normal tabular dataset and we just have to predict the probability of score of Team A or B depending on the position and velocity of ball and players , and hence people are using classification algorithms to win this challenge ?
", "rawMarkdown": "Thanks @aatiffraz and @ravi20076 , so this is a normal tabular dataset and we just have to predict the probability of score of Team A or B depending on the position and velocity of ball and players , and hence people are using classification algorithms to win this challenge ?", "votes": null }, { "id": "1968425", "postDate": "10/03/2022 04:47:41", "content": "Yes, that is right. You are right too, this table contains time series and panel information, this can also be used to generate a different approach. This problem can be solved in many ways.
", "rawMarkdown": "Yes, that is right. You are right too, this table contains time series and panel information, this can also be used to generate a different approach. This problem can be solved in many ways.", "votes": null } ], "comments": [ { "id": 1968012, "author_name": "aatiffraz", "author_url": "", "post_date": "10/02/2022 19:55:25", "content": "Hey @g9jiggy, it's because the test dataset is shuffled. Only the training dataset provides timestamps, sorted by game_num and event_id.
", "votes": null, "replies": [] }, { "id": 1968015, "author_name": "ravi20076", "author_url": "", "post_date": "10/02/2022 20:05:02", "content": "You may use such models for this dataset, this dataset contains temporal information along with much more. Also @aatiffraz has correctly pointed out that the test set is randomly shuffled leading to a loss of time-based continuity.
\nHope this helps @g9jiggy
Thanks @aatiffraz and @ravi20076 , so this is a normal tabular dataset and we just have to predict the probability of score of Team A or B depending on the position and velocity of ball and players , and hence people are using classification algorithms to win this challenge ?
", "votes": null, "replies": [] }, { "id": 1968425, "author_name": "ravi20076", "author_url": "", "post_date": "10/03/2022 04:47:41", "content": "Yes, that is right. You are right too, this table contains time series and panel information, this can also be used to generate a different approach. This problem can be solved in many ways.
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1967952": "Hey community i was going through this competition and people's notes and i am unable to understand few things\n-> Isn't this a time series dataset , if yes i am seeing people using Lightgbm classifier ? , i read code and discussion topics of this competition and just need a few starting pointers like how to approach this dataset what models will be best to use here", "1968012": "Hey @g9jiggy, it's because the test dataset is shuffled. Only the training dataset provides timestamps, sorted by game_num and event_id.", "1968015": "You may use such models for this dataset, this dataset contains temporal information along with much more. Also @aatiffraz has correctly pointed out that the test set is randomly shuffled leading to a loss of time-based continuity.\nHope this helps @g9jiggy", "1968344": "Thanks @aatiffraz and @ravi20076 , so this is a normal tabular dataset and we just have to predict the probability of score of Team A or B depending on the position and velocity of ball and players , and hence people are using classification algorithms to win this challenge ?", "1968425": "Yes, that is right. You are right too, this table contains time series and panel information, this can also be used to generate a different approach. This problem can be solved in many ways." }, "source": "meta" }