Hi all,\nfrom the datasets i see that the month_lag equals to 0 when date equals to 2/2018. So what is the target? calculate the transactions value for the next 2 months from the reference date? that means calculate the transactions for the dates 3/2018 to 4/2018?
", "messages": [ { "id": 429497, "postDate": "2018-11-29T00:11:28.047Z", "content": "Hi all,\nfrom the datasets i see that the month_lag equals to 0 when date equals to 2/2018. So what is the target? calculate the transactions value for the next 2 months from the reference date? that means calculate the transactions for the dates 3/2018 to 4/2018?
", "rawMarkdown": "Hi all,\nfrom the datasets i see that the month_lag equals to 0 when date equals to 2/2018. So what is the target? calculate the transactions value for the next 2 months from the reference date? that means calculate the transactions for the dates 3/2018 to 4/2018?", "votes": 8 }, { "id": 430923, "postDate": "2018-12-01T08:15:12.143Z", "content": "I believe that the intention of this competition is not to calculate a transaction value, but rather a \"Loyalty Score\". This is an internal business metric which takes into account both future spending and retention as main components. (Taken from https://www.kaggle.com/c/elo-merchant-category-recommendation/discussion/72993#430388)
\n\nTherefore, I think the main purpose is actually to find out if there is any underlying relationship between transactions behaviour and a customer's loyalty score. After knowing this relationship, ELO can predict a customer's loyalty towards ELO based on their historical transaction.
\n\nAs for month_lag, I think it is just to let you know how \"far\" away is this transaction from the reference date, which I believe is Feb 2018.
\n\nI'm not entirely sure why there would be a new merchant transaction and historical transaction file. I think it is just a way to add in more information, which may or may not be useful in finding out the relationships.
", "rawMarkdown": "I believe that the intention of this competition is not to calculate a transaction value, but rather a \"Loyalty Score\". This is an internal business metric which takes into account both future spending and retention as main components. (Taken from https://www.kaggle.com/c/elo-merchant-category-recommendation/discussion/72993#430388) \n\nTherefore, I think the main purpose is actually to find out if there is any underlying relationship between transactions behaviour and a customer's loyalty score. After knowing this relationship, ELO can predict a customer's loyalty towards ELO based on their historical transaction.\n\nAs for month_lag, I think it is just to let you know how \"far\" away is this transaction from the reference date, which I believe is Feb 2018.\n\nI'm not entirely sure why there would be a new merchant transaction and historical transaction file. I think it is just a way to add in more information, which may or may not be useful in finding out the relationships.", "votes": 3, "replies": [ { "id": 431858, "postDate": "2018-12-03T02:51:36.103Z", "content": "Thanks for sharing.\nMy issue with the reference date is that transactions take place BEFORE (historical transaction) and AFTER (new transactions) the reference date. So I'm not certain what's the point of the reference date - is it to infer the loyalty score on the reference period of Feb 18? But if so, then whats the point of having the new transaction data...
", "rawMarkdown": "Thanks for sharing.\nMy issue with the reference date is that transactions take place BEFORE (historical transaction) and AFTER (new transactions) the reference date. So I'm not certain what's the point of the reference date - is it to infer the loyalty score on the reference period of Feb 18? But if so, then whats the point of having the new transaction data... \n", "votes": 1 }, { "id": 443770, "postDate": "2018-12-22T11:46:36.040Z", "content": "It seems that historical transactions with loyalties ( 17, 7, -7, -9, -33, for example ) are not quite different.\nIs there any possibility that loyalty is not given by Elo, but by someone else, like bank.\nWhat Elo has are those historical transactions which come from cooperated merchants.\nBut those are not or just a small portion of data for judging loyalty by banks.\nElo wants to estimate and predict how banks giving loyalty.\nIn short, loyalties are not calculated from data given, but Elo wants to find out relationship between loyalty and transactions, or find out behavior of customers.\nGuessing...
", "rawMarkdown": "It seems that historical transactions with loyalties ( 17, 7, -7, -9, -33, for example ) are not quite different.\nIs there any possibility that loyalty is not given by Elo, but by someone else, like bank.\nWhat Elo has are those historical transactions which come from cooperated merchants.\nBut those are not or just a small portion of data for judging loyalty by banks.\nElo wants to estimate and predict how banks giving loyalty.\nIn short, loyalties are not calculated from data given, but Elo wants to find out relationship between loyalty and transactions, or find out behavior of customers.\nGuessing...", "votes": 2 }, { "id": 444546, "postDate": "2018-12-24T08:34:06.873Z", "content": "May different card_id has different reference date. Because I found that when month_lag == 0, the purchase_date are different.
maybe we should classify the outlier points into a different class.
I also don't understand the month_lag term found in both historical and new merchant transactions either. Anyone has a opinion?
", "rawMarkdown": "I also don't understand the month_lag term found in both historical and new merchant transactions either. Anyone has a opinion?", "votes": 2, "replies": [ { "id": 434420, "postDate": "2018-12-06T11:15:32.293Z", "content": "See this post www.shanelynn.ie/summarising-aggregation-and-grouping-data-in-python-pandas/amp\nI think that\nmonth_lag can be used for aggregating data easily.
", "rawMarkdown": "See this post www.shanelynn.ie/summarising-aggregation-and-grouping-data-in-python-pandas/amp\nI think that\nmonth_lag can be used for aggregating data easily. ", "votes": 1, "isDeleted": true } ] } ], "comments": [ { "id": 430923, "author_name": "Tee Ming Yi", "author_url": "", "post_date": "2018-12-01T08:15:12.143000", "content": "I believe that the intention of this competition is not to calculate a transaction value, but rather a \"Loyalty Score\". This is an internal business metric which takes into account both future spending and retention as main components. (Taken from https://www.kaggle.com/c/elo-merchant-category-recommendation/discussion/72993#430388)
\n\nTherefore, I think the main purpose is actually to find out if there is any underlying relationship between transactions behaviour and a customer's loyalty score. After knowing this relationship, ELO can predict a customer's loyalty towards ELO based on their historical transaction.
\n\nAs for month_lag, I think it is just to let you know how \"far\" away is this transaction from the reference date, which I believe is Feb 2018.
\n\nI'm not entirely sure why there would be a new merchant transaction and historical transaction file. I think it is just a way to add in more information, which may or may not be useful in finding out the relationships.
", "votes": 3, "replies": [ { "id": 431858, "author_name": "XLguy", "author_url": "", "post_date": "2018-12-03T02:51:36.103000", "content": "Thanks for sharing.\nMy issue with the reference date is that transactions take place BEFORE (historical transaction) and AFTER (new transactions) the reference date. So I'm not certain what's the point of the reference date - is it to infer the loyalty score on the reference period of Feb 18? But if so, then whats the point of having the new transaction data...
", "votes": 1, "replies": [] }, { "id": 443770, "author_name": "shihyung", "author_url": "", "post_date": "2018-12-22T11:46:36.040000", "content": "It seems that historical transactions with loyalties ( 17, 7, -7, -9, -33, for example ) are not quite different.\nIs there any possibility that loyalty is not given by Elo, but by someone else, like bank.\nWhat Elo has are those historical transactions which come from cooperated merchants.\nBut those are not or just a small portion of data for judging loyalty by banks.\nElo wants to estimate and predict how banks giving loyalty.\nIn short, loyalties are not calculated from data given, but Elo wants to find out relationship between loyalty and transactions, or find out behavior of customers.\nGuessing...
", "votes": 2, "replies": [] }, { "id": 444546, "author_name": "Manchoi", "author_url": "", "post_date": "2018-12-24T08:34:06.873000", "content": "May different card_id has different reference date. Because I found that when month_lag == 0, the purchase_date are different.
maybe we should classify the outlier points into a different class.
I also don't understand the month_lag term found in both historical and new merchant transactions either. Anyone has a opinion?
", "votes": 2, "replies": [ { "id": 434420, "author_name": "", "author_url": "", "post_date": "2018-12-06T11:15:32.293000", "content": "See this post www.shanelynn.ie/summarising-aggregation-and-grouping-data-in-python-pandas/amp\nI think that\nmonth_lag can be used for aggregating data easily.
", "votes": 1, "replies": [] } ] } ], "raw_markdown_by_id": { "429497": "Hi all,\nfrom the datasets i see that the month_lag equals to 0 when date equals to 2/2018. So what is the target? calculate the transactions value for the next 2 months from the reference date? that means calculate the transactions for the dates 3/2018 to 4/2018?", "430923": "I believe that the intention of this competition is not to calculate a transaction value, but rather a \"Loyalty Score\". This is an internal business metric which takes into account both future spending and retention as main components. (Taken from https://www.kaggle.com/c/elo-merchant-category-recommendation/discussion/72993#430388) \n\nTherefore, I think the main purpose is actually to find out if there is any underlying relationship between transactions behaviour and a customer's loyalty score. After knowing this relationship, ELO can predict a customer's loyalty towards ELO based on their historical transaction.\n\nAs for month_lag, I think it is just to let you know how \"far\" away is this transaction from the reference date, which I believe is Feb 2018.\n\nI'm not entirely sure why there would be a new merchant transaction and historical transaction file. I think it is just a way to add in more information, which may or may not be useful in finding out the relationships.", "430218": "I also don't understand the month_lag term found in both historical and new merchant transactions either. Anyone has a opinion?" } }