Hey guys, I may have a little more question. How is it different from the old credit risk model in term of stability
\n(Credit risk model vs Credit risk model Stability).
\nIs the new model stable because we just have to write more robust code for long term use? is that it?
\nI just want to know why it call stability (I know it for longterm use whatsoever but I just don't see the different between the old model and the new one)
\nWhy don't just we use old model --> change the input to fit this new data --> called it stability
\nThank you so much man
Hey guys, I may have a little more question. How is it different from the old credit risk model in term of stability
\n(Credit risk model vs Credit risk model Stability).
\nIs the new model stable because we just have to write more robust code for long term use? is that it?
\nI just want to know why it call stability (I know it for longterm use whatsoever but I just don't see the different between the old model and the new one)
\nWhy don't just we use old model --> change the input to fit this new data --> called it stability
\nThank you so much man
We need to make our solution more robust, which has nothing to do with the so-called model. To put it more bluntly, the evaluation metrics for the current competition have changed, so the designed solution will also change accordingly.😀
\nIn overview, we know:
\n\n", "rawMarkdown": "We need to make our solution more robust, which has nothing to do with the so-called model. To put it more bluntly, the evaluation metrics for the current competition have changed, so the designed solution will also change accordingly.😀\n\nIn overview, we know: \n> Currently, consumer finance providers use various statistical and machine learning methods to predict loan risk. These models are generally called scorecards. In the real world, clients' behaviors change constantly, so every scorecard must be updated regularly, which takes time. The scorecard's stability in the future is critical, as a sudden drop in performance means that loans will be issued to worse clients on average. The core of the issue is that loan providers aren't able to spot potential problems any sooner than the first due dates of those loans are observable. Given the time it takes to redevelop, validate, and implement the scorecard, stability is highly desirable. There is a trade-off between the stability of the model and its performance, and a balance must be reached before deployment.", "votes": null } ], "comments": [ { "id": 2765479, "author_name": "roger92", "author_url": "", "post_date": "04/21/2024 06:50:02", "content": "Currently, consumer finance providers use various statistical and machine learning methods to predict loan risk. These models are generally called scorecards. In the real world, clients' behaviors change constantly, so every scorecard must be updated regularly, which takes time. The scorecard's stability in the future is critical, as a sudden drop in performance means that loans will be issued to worse clients on average. The core of the issue is that loan providers aren't able to spot potential problems any sooner than the first due dates of those loans are observable. Given the time it takes to redevelop, validate, and implement the scorecard, stability is highly desirable. There is a trade-off between the stability of the model and its performance, and a balance must be reached before deployment.
\n
We need to make our solution more robust, which has nothing to do with the so-called model. To put it more bluntly, the evaluation metrics for the current competition have changed, so the designed solution will also change accordingly.😀
\nIn overview, we know:
\n\n", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "2752060": "Hey guys, I may have a little more question. How is it different from the old credit risk model in term of stability \n\n(Credit risk model vs Credit risk model Stability).\nIs the new model stable because we just have to write more robust code for long term use? is that it?\nI just want to know why it call stability (I know it for longterm use whatsoever but I just don't see the different between the old model and the new one) \nWhy don't just we use old model --> change the input to fit this new data --> called it stability\nThank you so much man", "2765479": "We need to make our solution more robust, which has nothing to do with the so-called model. To put it more bluntly, the evaluation metrics for the current competition have changed, so the designed solution will also change accordingly.😀\n\nIn overview, we know: \n> Currently, consumer finance providers use various statistical and machine learning methods to predict loan risk. These models are generally called scorecards. In the real world, clients' behaviors change constantly, so every scorecard must be updated regularly, which takes time. The scorecard's stability in the future is critical, as a sudden drop in performance means that loans will be issued to worse clients on average. The core of the issue is that loan providers aren't able to spot potential problems any sooner than the first due dates of those loans are observable. Given the time it takes to redevelop, validate, and implement the scorecard, stability is highly desirable. There is a trade-off between the stability of the model and its performance, and a balance must be reached before deployment." }, "source": "meta" }Currently, consumer finance providers use various statistical and machine learning methods to predict loan risk. These models are generally called scorecards. In the real world, clients' behaviors change constantly, so every scorecard must be updated regularly, which takes time. The scorecard's stability in the future is critical, as a sudden drop in performance means that loans will be issued to worse clients on average. The core of the issue is that loan providers aren't able to spot potential problems any sooner than the first due dates of those loans are observable. Given the time it takes to redevelop, validate, and implement the scorecard, stability is highly desirable. There is a trade-off between the stability of the model and its performance, and a balance must be reached before deployment.
\n