{
  "id": 227342,
  "title": "Post Processing in this Competition",
  "url": "/competitions/indoor-location-navigation/discussion/227342",
  "author_name": "Ravi Shah",
  "post_date": "2021-03-20T01:41:48.464000",
  "votes": 4,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Post processing is becoming a common and powerful tool in this competition.</p>\n<p><strong>What is Post Processing?</strong><br>\nPost Processing is essentially routines and processing you perform to the results of your model after your model has made its predictions. This is done to fine tune the results, and in this competition, your post processing can make a significant impact on your results.</p>\n<p><strong>Post Processing in this Competition:</strong><br>\nSo far the two main types of post processing in this competition are snap to grid and cost minimization (links at bottom). However, these are not the only forms of post processing I’ve heard of in this competition, and I am sure there is more to come. Wifi features have been the main features for models; however, other features are being widely used in post processing procedures.</p>\n<p><strong>Hyperparameter Tuning for Post Processing:</strong><br>\nJust like models, you can hyperparameter tune post processing techniques. Here are some hyperparameters to consider:<br>\nSnap to Grid: threshold<br>\nCost Minimization: alpha and beta</p>\n<p><strong>Links to Post Processing Kernels:</strong><br>\nCost minimization:<a href=\"url\" target=\"_blank\"> </a><a href=\"https://www.kaggle.com/saitodevel01/indoor-post-processing-by-cost-minimization\" target=\"_blank\">https://www.kaggle.com/saitodevel01/indoor-post-processing-by-cost-minimization</a> <br>\nSnap to Grid: <a href=\"url\" target=\"_blank\">https://www.kaggle.com/robikscube/indoor-navigation-snap-to-grid-post-processing</a> </p>\n<p>How much has post processing improved your results? What other post processing methods should we use?</p>",
  "messages": [
    {
      "id": 1245643,
      "postDate": "2021-03-20T01:41:48.463Z",
      "content": "<p>Post processing is becoming a common and powerful tool in this competition.</p>\n<p><strong>What is Post Processing?</strong><br>\nPost Processing is essentially routines and processing you perform to the results of your model after your model has made its predictions. This is done to fine tune the results, and in this competition, your post processing can make a significant impact on your results.</p>\n<p><strong>Post Processing in this Competition:</strong><br>\nSo far the two main types of post processing in this competition are snap to grid and cost minimization (links at bottom). However, these are not the only forms of post processing I’ve heard of in this competition, and I am sure there is more to come. Wifi features have been the main features for models; however, other features are being widely used in post processing procedures.</p>\n<p><strong>Hyperparameter Tuning for Post Processing:</strong><br>\nJust like models, you can hyperparameter tune post processing techniques. Here are some hyperparameters to consider:<br>\nSnap to Grid: threshold<br>\nCost Minimization: alpha and beta</p>\n<p><strong>Links to Post Processing Kernels:</strong><br>\nCost minimization:<a href=\"url\" target=\"_blank\"> </a><a href=\"https://www.kaggle.com/saitodevel01/indoor-post-processing-by-cost-minimization\" target=\"_blank\">https://www.kaggle.com/saitodevel01/indoor-post-processing-by-cost-minimization</a> <br>\nSnap to Grid: <a href=\"url\" target=\"_blank\">https://www.kaggle.com/robikscube/indoor-navigation-snap-to-grid-post-processing</a> </p>\n<p>How much has post processing improved your results? What other post processing methods should we use?</p>",
      "rawMarkdown": "Post processing is becoming a common and powerful tool in this competition.\n\n**What is Post Processing?**\nPost Processing is essentially routines and processing you perform to the results of your model after your model has made its predictions. This is done to fine tune the results, and in this competition, your post processing can make a significant impact on your results.\n\n**Post Processing in this Competition:**\nSo far the two main types of post processing in this competition are snap to grid and cost minimization (links at bottom). However, these are not the only forms of post processing I’ve heard of in this competition, and I am sure there is more to come. Wifi features have been the main features for models; however, other features are being widely used in post processing procedures.\n\n**Hyperparameter Tuning for Post Processing:**\nJust like models, you can hyperparameter tune post processing techniques. Here are some hyperparameters to consider:\nSnap to Grid: threshold\nCost Minimization: alpha and beta\n\n**Links to Post Processing Kernels:**\nCost minimization:[ https://www.kaggle.com/saitodevel01/indoor-post-processing-by-cost-minimization ](url)\nSnap to Grid: [https://www.kaggle.com/robikscube/indoor-navigation-snap-to-grid-post-processing](url) \n\nHow much has post processing improved your results? What other post processing methods should we use?",
      "votes": 4
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1245643": "Post processing is becoming a common and powerful tool in this competition.\n\n**What is Post Processing?**\nPost Processing is essentially routines and processing you perform to the results of your model after your model has made its predictions. This is done to fine tune the results, and in this competition, your post processing can make a significant impact on your results.\n\n**Post Processing in this Competition:**\nSo far the two main types of post processing in this competition are snap to grid and cost minimization (links at bottom). However, these are not the only forms of post processing I’ve heard of in this competition, and I am sure there is more to come. Wifi features have been the main features for models; however, other features are being widely used in post processing procedures.\n\n**Hyperparameter Tuning for Post Processing:**\nJust like models, you can hyperparameter tune post processing techniques. Here are some hyperparameters to consider:\nSnap to Grid: threshold\nCost Minimization: alpha and beta\n\n**Links to Post Processing Kernels:**\nCost minimization:[ https://www.kaggle.com/saitodevel01/indoor-post-processing-by-cost-minimization ](url)\nSnap to Grid: [https://www.kaggle.com/robikscube/indoor-navigation-snap-to-grid-post-processing](url) \n\nHow much has post processing improved your results? What other post processing methods should we use?"
  }
}