m new to kaggle competitions ,\nPlease let me know how does Private LB score calculate,i submit score for only Public test set,\nLatter how admins calculate the private LB score out of that .\n...Do people have to submit their model ??\nSuppose if do some preprocessing on the images and get good score out of that then latter when private lb score is calculated how it works how preprocessing will happen on private data set as it required for model to perform?
", "messages": [ { "id": "390375", "postDate": "09/20/2018 05:20:39", "content": "m new to kaggle competitions ,\nPlease let me know how does Private LB score calculate,i submit score for only Public test set,\nLatter how admins calculate the private LB score out of that .\n...Do people have to submit their model ??\nSuppose if do some preprocessing on the images and get good score out of that then latter when private lb score is calculated how it works how preprocessing will happen on private data set as it required for model to perform?
", "rawMarkdown": "m new to kaggle competitions ,\nPlease let me know how does Private LB score calculate,i submit score for only Public test set,\nLatter how admins calculate the private LB score out of that .\n...Do people have to submit their model ??\nSuppose if do some preprocessing on the images and get good score out of that then latter when private lb score is calculated how it works how preprocessing will happen on private data set as it required for model to perform?", "votes": null }, { "id": "390607", "postDate": "09/20/2018 14:01:50", "content": "In this challenge the private score is calculated on 85% of the test set and the public ranking score is calculated in 15% of the test set. About the submission of models each contest establishes how the rules will be. In general, only the winners must provide the models in order to receive the prizes. On preprocessing the general rule is that the process must be automated, so that it can be applied to new data. You should not use manual data preprocessing.
", "rawMarkdown": "In this challenge the private score is calculated on 85% of the test set and the public ranking score is calculated in 15% of the test set. About the submission of models each contest establishes how the rules will be. In general, only the winners must provide the models in order to receive the prizes. On preprocessing the general rule is that the process must be automated, so that it can be applied to new data. You should not use manual data preprocessing.", "votes": null }, { "id": "391612", "postDate": "09/22/2018 05:22:12", "content": "Thanks for info I know that part 85 15 pct split ......But how kaggle can determine private score based on just public test submission ,is private one also released and participants are asked to resubmit their score towards the end ?
", "rawMarkdown": "Thanks for info I know that part 85 15 pct split ......But how kaggle can determine private score based on just public test submission ,is private one also released and participants are asked to resubmit their score towards the end ?", "votes": null }, { "id": "391809", "postDate": "09/22/2018 13:19:32", "content": "When we submit our forecasts to 100% the test set the public and private scores are immediately calculated. This division of scores is necessary to avoid the exploitation of the information contained in the score to find the correct predictions.
", "rawMarkdown": "When we submit our forecasts to 100% the test set the public and private scores are immediately calculated. This division of scores is necessary to avoid the exploitation of the information contained in the score to find the correct predictions.", "votes": null } ], "comments": [ { "id": 390607, "author_name": "paulorzp", "author_url": "", "post_date": "09/20/2018 14:01:50", "content": "In this challenge the private score is calculated on 85% of the test set and the public ranking score is calculated in 15% of the test set. About the submission of models each contest establishes how the rules will be. In general, only the winners must provide the models in order to receive the prizes. On preprocessing the general rule is that the process must be automated, so that it can be applied to new data. You should not use manual data preprocessing.
", "votes": null, "replies": [ { "id": 391612, "author_name": "jaideepvalani", "author_url": "", "post_date": "09/22/2018 05:22:12", "content": "Thanks for info I know that part 85 15 pct split ......But how kaggle can determine private score based on just public test submission ,is private one also released and participants are asked to resubmit their score towards the end ?
", "votes": null, "replies": [] }, { "id": 391809, "author_name": "paulorzp", "author_url": "", "post_date": "09/22/2018 13:19:32", "content": "When we submit our forecasts to 100% the test set the public and private scores are immediately calculated. This division of scores is necessary to avoid the exploitation of the information contained in the score to find the correct predictions.
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "390375": "m new to kaggle competitions ,\nPlease let me know how does Private LB score calculate,i submit score for only Public test set,\nLatter how admins calculate the private LB score out of that .\n...Do people have to submit their model ??\nSuppose if do some preprocessing on the images and get good score out of that then latter when private lb score is calculated how it works how preprocessing will happen on private data set as it required for model to perform?", "390607": "In this challenge the private score is calculated on 85% of the test set and the public ranking score is calculated in 15% of the test set. About the submission of models each contest establishes how the rules will be. In general, only the winners must provide the models in order to receive the prizes. On preprocessing the general rule is that the process must be automated, so that it can be applied to new data. You should not use manual data preprocessing.", "391612": "Thanks for info I know that part 85 15 pct split ......But how kaggle can determine private score based on just public test submission ,is private one also released and participants are asked to resubmit their score towards the end ?", "391809": "When we submit our forecasts to 100% the test set the public and private scores are immediately calculated. This division of scores is necessary to avoid the exploitation of the information contained in the score to find the correct predictions." }, "source": "meta" }