Can anyone clarify column \"class\" in items file? Another moment is about the values for the column \"perishable\". Why does it have binary integer values instead of values either 1.25 and 1.0 as described in the description?
", "messages": [ { "id": "234778", "postDate": "10/24/2017 02:39:52", "content": "Can anyone clarify column \"class\" in items file? Another moment is about the values for the column \"perishable\". Why does it have binary integer values instead of values either 1.25 and 1.0 as described in the description?
", "rawMarkdown": "Can anyone clarify column \"class\" in items file? Another moment is about the values for the column \"perishable\". Why does it have binary integer values instead of values either 1.25 and 1.0 as described in the description?", "votes": null }, { "id": "234881", "postDate": "10/24/2017 08:52:54", "content": "Perishable means items which can not be stored for a long time. It's just an attribute of an item. Based on that attribute, the weights are selected for evaluation metric (1.25 or 1.0). \nIt makes some sense if you imagine consequences of demand prediction for a real store. You should penalize more for the errors in items, which can not live long in a store and will be debited if their demand prediction is higher than actual demand. That metric will reflect a business error value.
", "rawMarkdown": "Perishable means items which can not be stored for a long time. It's just an attribute of an item. Based on that attribute, the weights are selected for evaluation metric (1.25 or 1.0). \nIt makes some sense if you imagine consequences of demand prediction for a real store. You should penalize more for the errors in items, which can not live long in a store and will be debited if their demand prediction is higher than actual demand. That metric will reflect a business error value.", "votes": null }, { "id": "234886", "postDate": "10/24/2017 09:05:34", "content": "Thanks for the clarification! Do you have any comments on the column \"class\" in the same table?
", "rawMarkdown": "Thanks for the clarification! Do you have any comments on the column \"class\" in the same table?", "votes": null }, { "id": "234888", "postDate": "10/24/2017 09:07:32", "content": "Nope, this is a question for me also.
", "rawMarkdown": "Nope, this is a question for me also.", "votes": null }, { "id": "235288", "postDate": "10/25/2017 03:47:54", "content": "Family is a high-level description of the item's department and Class is a more refined category for each of the products.
", "rawMarkdown": "Family is a high-level description of the item's department and Class is a more refined category for each of the products.", "votes": null }, { "id": "235858", "postDate": "10/26/2017 06:53:06", "content": "Any more explanation/textual description on column \"Class\"?
", "rawMarkdown": "Any more explanation/textual description on column \"Class\"?", "votes": null }, { "id": "236239", "postDate": "10/26/2017 21:27:42", "content": "Family : Dairy / Class: Butter, Milk, Cheese and Ice Cream
", "rawMarkdown": "Family : Dairy / Class: Butter, Milk, Cheese and Ice Cream", "votes": null }, { "id": "236378", "postDate": "10/27/2017 06:48:55", "content": "Thanks for the clarification. Since they are just numbers, there is no way to understand whether something refers to Butter or Milk or Cheese, is it not? Will it not help to have them as text itself ? Logically speaking, frequency of buying milk will be different from buying Ice cream, so was wondering whether it would be good to have exact items. Or, was it kept deliberately this way for any other reasons?
", "rawMarkdown": "Thanks for the clarification. Since they are just numbers, there is no way to understand whether something refers to Butter or Milk or Cheese, is it not? Will it not help to have them as text itself ? Logically speaking, frequency of buying milk will be different from buying Ice cream, so was wondering whether it would be good to have exact items. Or, was it kept deliberately this way for any other reasons?", "votes": null }, { "id": "236489", "postDate": "10/27/2017 13:36:41", "content": "Items UPC, store names, store types and class types where deliberately anonymized for confidentiality reasons.
", "rawMarkdown": "Items UPC, store names, store types and class types where deliberately anonymized for confidentiality reasons.", "votes": null }, { "id": "237847", "postDate": "10/31/2017 05:17:09", "content": "Thanks for the clarification.
", "rawMarkdown": "Thanks for the clarification.", "votes": null } ], "comments": [ { "id": 234881, "author_name": "tsdaemon", "author_url": "", "post_date": "10/24/2017 08:52:54", "content": "Perishable means items which can not be stored for a long time. It's just an attribute of an item. Based on that attribute, the weights are selected for evaluation metric (1.25 or 1.0). \nIt makes some sense if you imagine consequences of demand prediction for a real store. You should penalize more for the errors in items, which can not live long in a store and will be debited if their demand prediction is higher than actual demand. That metric will reflect a business error value.
", "votes": null, "replies": [ { "id": 234886, "author_name": "yizenov", "author_url": "", "post_date": "10/24/2017 09:05:34", "content": "Thanks for the clarification! Do you have any comments on the column \"class\" in the same table?
", "votes": null, "replies": [] }, { "id": 234888, "author_name": "tsdaemon", "author_url": "", "post_date": "10/24/2017 09:07:32", "content": "Nope, this is a question for me also.
", "votes": null, "replies": [] } ] }, { "id": 235288, "author_name": "gyorgy", "author_url": "", "post_date": "10/25/2017 03:47:54", "content": "Family is a high-level description of the item's department and Class is a more refined category for each of the products.
", "votes": null, "replies": [] }, { "id": 235858, "author_name": "sambhavi", "author_url": "", "post_date": "10/26/2017 06:53:06", "content": "Any more explanation/textual description on column \"Class\"?
", "votes": null, "replies": [ { "id": 236239, "author_name": "gyorgy", "author_url": "", "post_date": "10/26/2017 21:27:42", "content": "Family : Dairy / Class: Butter, Milk, Cheese and Ice Cream
", "votes": null, "replies": [] } ] }, { "id": 236378, "author_name": "sambhavi", "author_url": "", "post_date": "10/27/2017 06:48:55", "content": "Thanks for the clarification. Since they are just numbers, there is no way to understand whether something refers to Butter or Milk or Cheese, is it not? Will it not help to have them as text itself ? Logically speaking, frequency of buying milk will be different from buying Ice cream, so was wondering whether it would be good to have exact items. Or, was it kept deliberately this way for any other reasons?
", "votes": null, "replies": [ { "id": 236489, "author_name": "gyorgy", "author_url": "", "post_date": "10/27/2017 13:36:41", "content": "Items UPC, store names, store types and class types where deliberately anonymized for confidentiality reasons.
", "votes": null, "replies": [] }, { "id": 237847, "author_name": "sambhavi", "author_url": "", "post_date": "10/31/2017 05:17:09", "content": "Thanks for the clarification.
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "234778": "Can anyone clarify column \"class\" in items file? Another moment is about the values for the column \"perishable\". Why does it have binary integer values instead of values either 1.25 and 1.0 as described in the description?", "234881": "Perishable means items which can not be stored for a long time. It's just an attribute of an item. Based on that attribute, the weights are selected for evaluation metric (1.25 or 1.0). \nIt makes some sense if you imagine consequences of demand prediction for a real store. You should penalize more for the errors in items, which can not live long in a store and will be debited if their demand prediction is higher than actual demand. That metric will reflect a business error value.", "234886": "Thanks for the clarification! Do you have any comments on the column \"class\" in the same table?", "234888": "Nope, this is a question for me also.", "235288": "Family is a high-level description of the item's department and Class is a more refined category for each of the products.", "235858": "Any more explanation/textual description on column \"Class\"?", "236239": "Family : Dairy / Class: Butter, Milk, Cheese and Ice Cream", "236378": "Thanks for the clarification. Since they are just numbers, there is no way to understand whether something refers to Butter or Milk or Cheese, is it not? Will it not help to have them as text itself ? Logically speaking, frequency of buying milk will be different from buying Ice cream, so was wondering whether it would be good to have exact items. Or, was it kept deliberately this way for any other reasons?", "236489": "Items UPC, store names, store types and class types where deliberately anonymized for confidentiality reasons.", "237847": "Thanks for the clarification." }, "source": "meta" }