Teaching material for 'G2net - Machine Learning and Signal processing for Time Series Analysis' held at the University of Malta, Valletta Campus, Malta, 9-13 March 2020.
\nTeaching material for 'G2net - Machine Learning and Signal processing for Time Series Analysis' held at the University of Malta, Valletta Campus, Malta, 9-13 March 2020.
\nHey, if anyone needs help with the content of the training school, please feel free to reach out.
", "rawMarkdown": "Hey, if anyone needs help with the content of the training school, please feel free to reach out.", "votes": null }, { "id": "1373220", "postDate": "07/02/2021 10:19:16", "content": "Good day,
\nFinally I see the right setup for the problem.
\nI was researching and know from reading the material in MATLAB that you needed to
\nHave two CNNs and a soft ReLu output and a Bayesian network for prediction… (this is an oversimplification of the setup and much is left out just for the posting).
\nGreat work thanks
\nP.S. can MATLAB be use and imported (converted) in to Kaggle?
\nB.
\ncaptbullett
Kaggle supports majorly Python and R. We have to convert the code to either to use in notebooks
", "rawMarkdown": "Kaggle supports majorly Python and R. We have to convert the code to either to use in notebooks", "votes": null }, { "id": "1373247", "postDate": "07/02/2021 10:52:35", "content": "Thank you,
\nYes I understand most of us forget that R us for this type of ML and it’s hard to implement sometimes but once you’re passed the learning curve it easy as py.
\nThanks B.
\ncaptbullett
Thanks of sharing. I'm doubtful if I'd be able to get through all of this within the first few weeks of the competition, but it's a nice bookmark to have 😊.
", "rawMarkdown": "Thanks of sharing. I'm doubtful if I'd be able to get through all of this within the first few weeks of the competition, but it's a nice bookmark to have 😊.", "votes": null }, { "id": "1375786", "postDate": "07/04/2021 13:37:57", "content": "Thanks for the resources but I think lecture links are redirecting back to the discussion forum
", "rawMarkdown": "Thanks for the resources but I think lecture links are redirecting back to the discussion forum", "votes": null }, { "id": "1380517", "postDate": "07/08/2021 06:29:44", "content": "Thanks for pointing it out. I have fixed this.
", "rawMarkdown": "Thanks for pointing it out. I have fixed this.", "votes": null }, { "id": "1381546", "postDate": "07/09/2021 05:09:23", "content": "I see many people taking the wave -> image -> model approach which does not consider the SNR explanation given by the authors in the description. The reason being this way always gives better results than taking the signal processing route. Is this what everyone is doing? Or is there another \"traditional\" method that you suggest based on the material @shaz13 shared?
", "rawMarkdown": "I see many people taking the wave -> image -> model approach which does not consider the SNR explanation given by the authors in the description. The reason being this way always gives better results than taking the signal processing route. Is this what everyone is doing? Or is there another \"traditional\" method that you suggest based on the material @shaz13 shared?", "votes": null }, { "id": "1381944", "postDate": "07/09/2021 11:23:39", "content": "I also noticed SNR is mentioned as important information by host, but no one used it from discussion. Maybe we can use CNN as main classifier and use SNR as post-process factor. But one question is that how to calculate accurate SNR in Python, one simple way is to use signal's mean and std for calculating SNR but looks this way is not so accurate. Any idea of better SNR calucation? @siddharthpatel45
", "rawMarkdown": "I also noticed SNR is mentioned as important information by host, but no one used it from discussion. Maybe we can use CNN as main classifier and use SNR as post-process factor. But one question is that how to calculate accurate SNR in Python, one simple way is to use signal's mean and std for calculating SNR but looks this way is not so accurate. Any idea of better SNR calucation? @siddharthpatel45", "votes": null }, { "id": "1382513", "postDate": "07/10/2021 00:56:45", "content": "thanks for the suggestion.
\nAbout the SNR, actually I don't have much signal processing background so was hoping someone here would point in the direction to look for.
You cannot calculate SNR with only one signal. You need true signal and noise as separate signals. You can approximate it using some heuristics like in https://www.researchgate.net/post/Does_anyone_know_how_to_get_SNR_signal_to_noise_ratio_for_real_data but not sure how applicable those are to this case
", "rawMarkdown": "You cannot calculate SNR with only one signal. You need true signal and noise as separate signals. You can approximate it using some heuristics like in https://www.researchgate.net/post/Does_anyone_know_how_to_get_SNR_signal_to_noise_ratio_for_real_data but not sure how applicable those are to this case", "votes": null }, { "id": "1400870", "postDate": "07/26/2021 17:02:08", "content": "Video recordings would be great!
", "rawMarkdown": "Video recordings would be great!", "votes": null }, { "id": "1488037", "postDate": "08/24/2021 03:26:13", "content": "Not all lectures have recording, but some do. You can find details in my Git here https://github.com/zerafachris/g2net_2nd_training_school_malta_mar_2020
", "rawMarkdown": "Not all lectures have recording, but some do. You can find details in my Git here https://github.com/zerafachris/g2net_2nd_training_school_malta_mar_2020", "votes": null }, { "id": "1492469", "postDate": "08/27/2021 07:42:41", "content": "Thank you for sharing such detailed material for the tasks.
", "rawMarkdown": "Thank you for sharing such detailed material for the tasks.", "votes": null }, { "id": "1496732", "postDate": "08/30/2021 15:34:10", "content": "Great Work,thank you
", "rawMarkdown": "Great Work,thank you", "votes": null }, { "id": "1561010", "postDate": "10/27/2021 09:09:21", "content": "Hey All,
\nThank you all for taking part in our competition. The participation has been overwhelmingly positive. We are currently conducting a survey to gauge the demographic and outreach achieved. Kindly spare 2min and fill in this survey https://forms.gle/QP9L16niPexozyhu5.
\nThank you all,
\nRegards,
\nChris
Hey, if anyone needs help with the content of the training school, please feel free to reach out.
", "votes": null, "replies": [ { "id": 1400870, "author_name": "bluesky314", "author_url": "", "post_date": "07/26/2021 17:02:08", "content": "Video recordings would be great!
", "votes": null, "replies": [] }, { "id": 1488037, "author_name": "zerafachris", "author_url": "", "post_date": "08/24/2021 03:26:13", "content": "Not all lectures have recording, but some do. You can find details in my Git here https://github.com/zerafachris/g2net_2nd_training_school_malta_mar_2020
", "votes": null, "replies": [] } ] }, { "id": 1373220, "author_name": "captbullett", "author_url": "", "post_date": "07/02/2021 10:19:16", "content": "Good day,
\nFinally I see the right setup for the problem.
\nI was researching and know from reading the material in MATLAB that you needed to
\nHave two CNNs and a soft ReLu output and a Bayesian network for prediction… (this is an oversimplification of the setup and much is left out just for the posting).
\nGreat work thanks
\nP.S. can MATLAB be use and imported (converted) in to Kaggle?
\nB.
\ncaptbullett
Kaggle supports majorly Python and R. We have to convert the code to either to use in notebooks
", "votes": null, "replies": [] } ] }, { "id": 1373247, "author_name": "captbullett", "author_url": "", "post_date": "07/02/2021 10:52:35", "content": "Thank you,
\nYes I understand most of us forget that R us for this type of ML and it’s hard to implement sometimes but once you’re passed the learning curve it easy as py.
\nThanks B.
\ncaptbullett
Thanks of sharing. I'm doubtful if I'd be able to get through all of this within the first few weeks of the competition, but it's a nice bookmark to have 😊.
", "votes": null, "replies": [] }, { "id": 1375786, "author_name": "mrigendraagrawal", "author_url": "", "post_date": "07/04/2021 13:37:57", "content": "Thanks for the resources but I think lecture links are redirecting back to the discussion forum
", "votes": null, "replies": [ { "id": 1380517, "author_name": "shaz13", "author_url": "", "post_date": "07/08/2021 06:29:44", "content": "Thanks for pointing it out. I have fixed this.
", "votes": null, "replies": [] } ] }, { "id": 1381546, "author_name": "siddharthpatel45", "author_url": "", "post_date": "07/09/2021 05:09:23", "content": "I see many people taking the wave -> image -> model approach which does not consider the SNR explanation given by the authors in the description. The reason being this way always gives better results than taking the signal processing route. Is this what everyone is doing? Or is there another \"traditional\" method that you suggest based on the material @shaz13 shared?
", "votes": null, "replies": [ { "id": 1381944, "author_name": "superchenhao", "author_url": "", "post_date": "07/09/2021 11:23:39", "content": "I also noticed SNR is mentioned as important information by host, but no one used it from discussion. Maybe we can use CNN as main classifier and use SNR as post-process factor. But one question is that how to calculate accurate SNR in Python, one simple way is to use signal's mean and std for calculating SNR but looks this way is not so accurate. Any idea of better SNR calucation? @siddharthpatel45
", "votes": null, "replies": [] }, { "id": 1382513, "author_name": "siddharthpatel45", "author_url": "", "post_date": "07/10/2021 00:56:45", "content": "thanks for the suggestion.
\nAbout the SNR, actually I don't have much signal processing background so was hoping someone here would point in the direction to look for.
You cannot calculate SNR with only one signal. You need true signal and noise as separate signals. You can approximate it using some heuristics like in https://www.researchgate.net/post/Does_anyone_know_how_to_get_SNR_signal_to_noise_ratio_for_real_data but not sure how applicable those are to this case
", "votes": null, "replies": [] } ] }, { "id": 1492469, "author_name": "sauravsolanki", "author_url": "", "post_date": "08/27/2021 07:42:41", "content": "Thank you for sharing such detailed material for the tasks.
", "votes": null, "replies": [] }, { "id": 1496732, "author_name": "hanson0910", "author_url": "", "post_date": "08/30/2021 15:34:10", "content": "Great Work,thank you
", "votes": null, "replies": [] }, { "id": 1561010, "author_name": "zerafachris", "author_url": "", "post_date": "10/27/2021 09:09:21", "content": "Hey All,
\nThank you all for taking part in our competition. The participation has been overwhelmingly positive. We are currently conducting a survey to gauge the demographic and outreach achieved. Kindly spare 2min and fill in this survey https://forms.gle/QP9L16niPexozyhu5.
\nThank you all,
\nRegards,
\nChris