Hello all,\non the file merchant.csv, it is supposed to have some attributes per unique mechant id. But, on the merchant 'M_ID_1802942aaf' we have 4 rows (there are also other merchants with duplicate records).
\n\nThese mercahnts have different merchant_group_id, most_recent_purchases_range and lagged features.
\n\nWhat are we supposed to do with these mercahnts?
", "messages": [ { "id": "430109", "postDate": "11/29/2018 21:14:05", "content": "Hello all,\non the file merchant.csv, it is supposed to have some attributes per unique mechant id. But, on the merchant 'M_ID_1802942aaf' we have 4 rows (there are also other merchants with duplicate records).
\n\nThese mercahnts have different merchant_group_id, most_recent_purchases_range and lagged features.
\n\nWhat are we supposed to do with these mercahnts?
", "rawMarkdown": "Hello all,\non the file merchant.csv, it is supposed to have some attributes per unique mechant id. But, on the merchant 'M_ID_1802942aaf' we have 4 rows (there are also other merchants with duplicate records). \n\nThese mercahnts have different merchant_group_id, most_recent_purchases_range and lagged features.\n\nWhat are we supposed to do with these mercahnts?", "votes": null }, { "id": "430158", "postDate": "11/29/2018 23:22:57", "content": "The entire competition is fundamentally flawed. See my remarks here:
\n\nhttps://www.kaggle.com/c/elo-merchant-category-recommendation/discussion/73113
\n\nBasically, just do what everyone else is. Randomly cram different columns into different, existing machine learning libraries and see if you can beat everyone else. That's all this company is going to get for 50 grand.
", "rawMarkdown": "The entire competition is fundamentally flawed. See my remarks here:\n\nhttps://www.kaggle.com/c/elo-merchant-category-recommendation/discussion/73113\n\nBasically, just do what everyone else is. Randomly cram different columns into different, existing machine learning libraries and see if you can beat everyone else. That's all this company is going to get for 50 grand.", "votes": null }, { "id": "430300", "postDate": "11/30/2018 06:39:14", "content": "Well, unfortunately this sounds true
", "rawMarkdown": "Well, unfortunately this sounds true", "votes": null }, { "id": "430484", "postDate": "11/30/2018 13:08:08", "content": "I will rather think that the goal of this competition is to understand these cryptic things.
", "rawMarkdown": "I will rather think that the goal of this competition is to understand these cryptic things.", "votes": null }, { "id": "430517", "postDate": "11/30/2018 13:59:11", "content": "*I think it is correct. Merchants can belong to different merchant groups, e.g., a restaurant can also sell groceries. Therefore, the merchant will belong to different merchant groups. Based on this way of thinking, depending on what the card_id bought, the most recentpurchases_range will also change.*
", "rawMarkdown": "*I think it is correct. Merchants can belong to different merchant groups, e.g., a restaurant can also sell groceries. Therefore, the merchant will belong to different merchant groups. Based on this way of thinking, depending on what the card_id bought, the most recentpurchases_range will also change.*", "votes": null }, { "id": "442405", "postDate": "12/19/2018 23:00:25", "content": "yes, I can understand multiple entries due to merchant category and it can happen in two cases - \n1 - A merchant can have a primary category and secondary category. Take gas stations for instance, their primary line of business will be Gas station but there is also a secondary category which is retail store inside gas station. \n2 - A merchant has moved from one category to another recently.
\n\nBut, there are multiple entries for same merchant for due to most_recent_sales_range, most_recent_purchases_range etc. Like 1 records says E and other says D. Isnt most_recent_sales_range/most_recent_purchases_range is strictly supposed to be the MOST recent and there should be only one entry with most recent.
", "rawMarkdown": "yes, I can understand multiple entries due to merchant category and it can happen in two cases - \n1 - A merchant can have a primary category and secondary category. Take gas stations for instance, their primary line of business will be Gas station but there is also a secondary category which is retail store inside gas station. \n2 - A merchant has moved from one category to another recently. \n\nBut, there are multiple entries for same merchant for due to most_recent_sales_range, most_recent_purchases_range etc. Like 1 records says E and other says D. Isnt most_recent_sales_range/most_recent_purchases_range is strictly supposed to be the MOST recent and there should be only one entry with most recent.", "votes": null }, { "id": "442440", "postDate": "12/20/2018 01:00:14", "content": "I just dropped the duplicates and kept the first instance, when i was using merchant data.
", "rawMarkdown": "I just dropped the duplicates and kept the first instance, when i was using merchant data.", "votes": null }, { "id": "442575", "postDate": "12/20/2018 06:36:14", "content": "did it help? I gain nothing...
", "rawMarkdown": "did it help? I gain nothing...", "votes": null }, { "id": "442832", "postDate": "12/20/2018 15:27:58", "content": "No help from merchant data so far.
", "rawMarkdown": "No help from merchant data so far.", "votes": null } ], "comments": [ { "id": 430158, "author_name": "technikempire", "author_url": "", "post_date": "11/29/2018 23:22:57", "content": "The entire competition is fundamentally flawed. See my remarks here:
\n\nhttps://www.kaggle.com/c/elo-merchant-category-recommendation/discussion/73113
\n\nBasically, just do what everyone else is. Randomly cram different columns into different, existing machine learning libraries and see if you can beat everyone else. That's all this company is going to get for 50 grand.
", "votes": null, "replies": [ { "id": 430300, "author_name": "dkaraflos", "author_url": "", "post_date": "11/30/2018 06:39:14", "content": "Well, unfortunately this sounds true
", "votes": null, "replies": [] } ] }, { "id": 430484, "author_name": "chambox", "author_url": "", "post_date": "11/30/2018 13:08:08", "content": "I will rather think that the goal of this competition is to understand these cryptic things.
", "votes": null, "replies": [] }, { "id": 430517, "author_name": "chambox", "author_url": "", "post_date": "11/30/2018 13:59:11", "content": "*I think it is correct. Merchants can belong to different merchant groups, e.g., a restaurant can also sell groceries. Therefore, the merchant will belong to different merchant groups. Based on this way of thinking, depending on what the card_id bought, the most recentpurchases_range will also change.*
", "votes": null, "replies": [] }, { "id": 442405, "author_name": "mrrc18", "author_url": "", "post_date": "12/19/2018 23:00:25", "content": "yes, I can understand multiple entries due to merchant category and it can happen in two cases - \n1 - A merchant can have a primary category and secondary category. Take gas stations for instance, their primary line of business will be Gas station but there is also a secondary category which is retail store inside gas station. \n2 - A merchant has moved from one category to another recently.
\n\nBut, there are multiple entries for same merchant for due to most_recent_sales_range, most_recent_purchases_range etc. Like 1 records says E and other says D. Isnt most_recent_sales_range/most_recent_purchases_range is strictly supposed to be the MOST recent and there should be only one entry with most recent.
", "votes": null, "replies": [] }, { "id": 442440, "author_name": "harshit92", "author_url": "", "post_date": "12/20/2018 01:00:14", "content": "I just dropped the duplicates and kept the first instance, when i was using merchant data.
", "votes": null, "replies": [ { "id": 442575, "author_name": "marcuslin", "author_url": "", "post_date": "12/20/2018 06:36:14", "content": "did it help? I gain nothing...
", "votes": null, "replies": [] }, { "id": 442832, "author_name": "harshit92", "author_url": "", "post_date": "12/20/2018 15:27:58", "content": "No help from merchant data so far.
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "430109": "Hello all,\non the file merchant.csv, it is supposed to have some attributes per unique mechant id. But, on the merchant 'M_ID_1802942aaf' we have 4 rows (there are also other merchants with duplicate records). \n\nThese mercahnts have different merchant_group_id, most_recent_purchases_range and lagged features.\n\nWhat are we supposed to do with these mercahnts?", "430158": "The entire competition is fundamentally flawed. See my remarks here:\n\nhttps://www.kaggle.com/c/elo-merchant-category-recommendation/discussion/73113\n\nBasically, just do what everyone else is. Randomly cram different columns into different, existing machine learning libraries and see if you can beat everyone else. That's all this company is going to get for 50 grand.", "430300": "Well, unfortunately this sounds true", "430484": "I will rather think that the goal of this competition is to understand these cryptic things.", "430517": "*I think it is correct. Merchants can belong to different merchant groups, e.g., a restaurant can also sell groceries. Therefore, the merchant will belong to different merchant groups. Based on this way of thinking, depending on what the card_id bought, the most recentpurchases_range will also change.*", "442405": "yes, I can understand multiple entries due to merchant category and it can happen in two cases - \n1 - A merchant can have a primary category and secondary category. Take gas stations for instance, their primary line of business will be Gas station but there is also a secondary category which is retail store inside gas station. \n2 - A merchant has moved from one category to another recently. \n\nBut, there are multiple entries for same merchant for due to most_recent_sales_range, most_recent_purchases_range etc. Like 1 records says E and other says D. Isnt most_recent_sales_range/most_recent_purchases_range is strictly supposed to be the MOST recent and there should be only one entry with most recent.", "442440": "I just dropped the duplicates and kept the first instance, when i was using merchant data.", "442575": "did it help? I gain nothing...", "442832": "No help from merchant data so far." }, "source": "meta" }