Hello all, I continue my discussion on online machine learning with the below python library references-
\n1. Creme
\na. The below resources may be of use to onboard to this library-
\nhttps://analyticsindiamag.com/how-to-learn-from-streaming-data-with-creme-in-python/ - this is a good starter article to learn the tenets of this package
\nb. https://pypi.org/project/creme/ - this is the official page of the package
\nc. https://pyimagesearch.com/2019/06/17/online-incremental-learning-with-keras-and-creme/ - this is a bit dated but is still useful
2. Scikit- multiflow
\nThe official documentation page is as follows- https://scikit-multiflow.readthedocs.io/en/stable/index.html
\nIt facilitates online learning algorithms, consists of several scikit-learn like structures and also is compatible with a pipeline architecture, facilitating drift detection too. It facilitates the below learning methods-
\na. Anomaly detection
\nb. Lazy learning
\nc. Bayes methods
\nd. Ensembles
\ne. Neural Networks
\nf. Rules and tree methods
3. River
\nThe below article introduces one to the library- https://towardsdatascience.com/river-the-best-python-library-for-online-machine-learning-56bf6f71a403
\nThis is a newer release that concocts the best of creme and scikit-multiflow and is specifically suited for online learning algorithms and pipelines. It provides effective drift detection and is quite suitable for imbalances in data too. The below links are useful herewith-
\na. https://github.com/online-ml/river
\nb. https://riverml.xyz/0.13.0/
\nc. https://analyticsindiamag.com/a-guide-to-river-a-python-tool-for-online-learning/ - introduces one to the package quite well
\nd. https://www.youtube.com/watch?v=1gGrb7Vpe80 - this video is quite good to onboard onto the library
All the best!!
", "messages": [ { "id": "1965146", "postDate": "10/01/2022 07:08:03", "content": "Hello all, I continue my discussion on online machine learning with the below python library references-
\n1. Creme
\na. The below resources may be of use to onboard to this library-
\nhttps://analyticsindiamag.com/how-to-learn-from-streaming-data-with-creme-in-python/ - this is a good starter article to learn the tenets of this package
\nb. https://pypi.org/project/creme/ - this is the official page of the package
\nc. https://pyimagesearch.com/2019/06/17/online-incremental-learning-with-keras-and-creme/ - this is a bit dated but is still useful
2. Scikit- multiflow
\nThe official documentation page is as follows- https://scikit-multiflow.readthedocs.io/en/stable/index.html
\nIt facilitates online learning algorithms, consists of several scikit-learn like structures and also is compatible with a pipeline architecture, facilitating drift detection too. It facilitates the below learning methods-
\na. Anomaly detection
\nb. Lazy learning
\nc. Bayes methods
\nd. Ensembles
\ne. Neural Networks
\nf. Rules and tree methods
3. River
\nThe below article introduces one to the library- https://towardsdatascience.com/river-the-best-python-library-for-online-machine-learning-56bf6f71a403
\nThis is a newer release that concocts the best of creme and scikit-multiflow and is specifically suited for online learning algorithms and pipelines. It provides effective drift detection and is quite suitable for imbalances in data too. The below links are useful herewith-
\na. https://github.com/online-ml/river
\nb. https://riverml.xyz/0.13.0/
\nc. https://analyticsindiamag.com/a-guide-to-river-a-python-tool-for-online-learning/ - introduces one to the package quite well
\nd. https://www.youtube.com/watch?v=1gGrb7Vpe80 - this video is quite good to onboard onto the library
All the best!!
", "rawMarkdown": "Hello all, I continue my discussion on online machine learning with the below python library references- \n\n**1. Creme**\na. The below resources may be of use to onboard to this library- \nhttps://analyticsindiamag.com/how-to-learn-from-streaming-data-with-creme-in-python/ - this is a good starter article to learn the tenets of this package\nb. https://pypi.org/project/creme/ - this is the official page of the package\nc. https://pyimagesearch.com/2019/06/17/online-incremental-learning-with-keras-and-creme/ - this is a bit dated but is still useful\n\n**2. Scikit- multiflow**\nThe official documentation page is as follows- https://scikit-multiflow.readthedocs.io/en/stable/index.html\nIt facilitates online learning algorithms, consists of several scikit-learn like structures and also is compatible with a pipeline architecture, facilitating drift detection too. It facilitates the below learning methods-\na. Anomaly detection\nb. Lazy learning \nc. Bayes methods\nd. Ensembles\ne. Neural Networks\nf. Rules and tree methods\n\n**3. River**\nThe below article introduces one to the library- https://towardsdatascience.com/river-the-best-python-library-for-online-machine-learning-56bf6f71a403\nThis is a newer release that concocts the best of creme and scikit-multiflow and is specifically suited for online learning algorithms and pipelines. It provides effective drift detection and is quite suitable for imbalances in data too. The below links are useful herewith- \na. https://github.com/online-ml/river\nb. https://riverml.xyz/0.13.0/\nc. https://analyticsindiamag.com/a-guide-to-river-a-python-tool-for-online-learning/ - introduces one to the package quite well\nd. https://www.youtube.com/watch?v=1gGrb7Vpe80 - this video is quite good to onboard onto the library\n\nAll the best!!", "votes": null }, { "id": "1966110", "postDate": "10/01/2022 18:02:29", "content": "Hi @ravi20076 these seems great, probably River is combination of the first 2.
\nFrom my understanding you need to use this API somewhere continuously.
\nHave you used this in any projects?
\nCan you give few points on architecture?
My job role does not require us to use these types of models, hence this is a new area for me. Hoping to learn lots of new model techniques from the month's assignment @abrafey
", "rawMarkdown": "My job role does not require us to use these types of models, hence this is a new area for me. Hoping to learn lots of new model techniques from the month's assignment @abrafey", "votes": null }, { "id": "1968334", "postDate": "10/03/2022 04:05:27", "content": "Thank you! Great information
", "rawMarkdown": "Thank you! Great information", "votes": null }, { "id": "1968896", "postDate": "10/03/2022 09:08:14", "content": "Welcome! Hoping this helps!!
", "rawMarkdown": "Welcome! Hoping this helps!!", "votes": null }, { "id": "1969015", "postDate": "10/03/2022 09:55:59", "content": "Great share @ravi20076
", "rawMarkdown": "Great share @ravi20076", "votes": null }, { "id": "1969814", "postDate": "10/03/2022 17:43:37", "content": "Most welcome @kinnerakiran, hope this helps with the competition!
", "rawMarkdown": "Most welcome @kinnerakiran, hope this helps with the competition!", "votes": null }, { "id": "1971394", "postDate": "10/04/2022 15:19:12", "content": "Great work!! @ravi20076
", "rawMarkdown": "Great work!! @ravi20076", "votes": null }, { "id": "1979589", "postDate": "10/09/2022 15:06:40", "content": "Many thanks, I am glad if this helped you @ananyasingh008
", "rawMarkdown": "Many thanks, I am glad if this helped you @ananyasingh008", "votes": null } ], "comments": [ { "id": 1966110, "author_name": "abrafey", "author_url": "", "post_date": "10/01/2022 18:02:29", "content": "Hi @ravi20076 these seems great, probably River is combination of the first 2.
\nFrom my understanding you need to use this API somewhere continuously.
\nHave you used this in any projects?
\nCan you give few points on architecture?
My job role does not require us to use these types of models, hence this is a new area for me. Hoping to learn lots of new model techniques from the month's assignment @abrafey
", "votes": null, "replies": [] } ] }, { "id": 1968334, "author_name": "willcramptonn", "author_url": "", "post_date": "10/03/2022 04:05:27", "content": "Thank you! Great information
", "votes": null, "replies": [ { "id": 1968896, "author_name": "ravi20076", "author_url": "", "post_date": "10/03/2022 09:08:14", "content": "Welcome! Hoping this helps!!
", "votes": null, "replies": [] } ] }, { "id": 1969015, "author_name": "kinnerakiran", "author_url": "", "post_date": "10/03/2022 09:55:59", "content": "Great share @ravi20076
", "votes": null, "replies": [ { "id": 1969814, "author_name": "ravi20076", "author_url": "", "post_date": "10/03/2022 17:43:37", "content": "Most welcome @kinnerakiran, hope this helps with the competition!
", "votes": null, "replies": [] } ] }, { "id": 1971394, "author_name": "ananyasingh008", "author_url": "", "post_date": "10/04/2022 15:19:12", "content": "Great work!! @ravi20076
", "votes": null, "replies": [ { "id": 1979589, "author_name": "ravi20076", "author_url": "", "post_date": "10/09/2022 15:06:40", "content": "Many thanks, I am glad if this helped you @ananyasingh008
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1965146": "Hello all, I continue my discussion on online machine learning with the below python library references- \n\n**1. Creme**\na. The below resources may be of use to onboard to this library- \nhttps://analyticsindiamag.com/how-to-learn-from-streaming-data-with-creme-in-python/ - this is a good starter article to learn the tenets of this package\nb. https://pypi.org/project/creme/ - this is the official page of the package\nc. https://pyimagesearch.com/2019/06/17/online-incremental-learning-with-keras-and-creme/ - this is a bit dated but is still useful\n\n**2. Scikit- multiflow**\nThe official documentation page is as follows- https://scikit-multiflow.readthedocs.io/en/stable/index.html\nIt facilitates online learning algorithms, consists of several scikit-learn like structures and also is compatible with a pipeline architecture, facilitating drift detection too. It facilitates the below learning methods-\na. Anomaly detection\nb. Lazy learning \nc. Bayes methods\nd. Ensembles\ne. Neural Networks\nf. Rules and tree methods\n\n**3. River**\nThe below article introduces one to the library- https://towardsdatascience.com/river-the-best-python-library-for-online-machine-learning-56bf6f71a403\nThis is a newer release that concocts the best of creme and scikit-multiflow and is specifically suited for online learning algorithms and pipelines. It provides effective drift detection and is quite suitable for imbalances in data too. The below links are useful herewith- \na. https://github.com/online-ml/river\nb. https://riverml.xyz/0.13.0/\nc. https://analyticsindiamag.com/a-guide-to-river-a-python-tool-for-online-learning/ - introduces one to the package quite well\nd. https://www.youtube.com/watch?v=1gGrb7Vpe80 - this video is quite good to onboard onto the library\n\nAll the best!!", "1966110": "Hi @ravi20076 these seems great, probably River is combination of the first 2.\nFrom my understanding you need to use this API somewhere continuously.\nHave you used this in any projects?\nCan you give few points on architecture?", "1966879": "My job role does not require us to use these types of models, hence this is a new area for me. Hoping to learn lots of new model techniques from the month's assignment @abrafey", "1968334": "Thank you! Great information", "1968896": "Welcome! Hoping this helps!!", "1969015": "Great share @ravi20076", "1969814": "Most welcome @kinnerakiran, hope this helps with the competition!", "1971394": "Great work!! @ravi20076", "1979589": "Many thanks, I am glad if this helped you @ananyasingh008" }, "source": "meta" }