Re: Does Weka do assumption check?

"Shi, Catherine" <[email protected]> Sat, 26 Aug 2023 15:15:09 +0000
Newsgroups gmane.comp.ai.weka
Message-ID <MN2PR01MB5438106ED71EA62F621A9B42EFE2A@MN2PR01MB5438.prod.exchangelabs.com>
Thank you for your input on my question! And yes, I’m not doing pre-checking. As you understand, I would like to build a linear regression model first, and then check for the assumptions based on residuals and see if there’s any lack of fit or any unreliable results from the model. Basically just model diagnostic things. Sometimes I may need to add other forms of the variables for a more reliable result. I could use a regression tree or random forest kind of thing as well for the same task. Thank you for pointing this out!

From: Michael Hall <[email protected]>
Date: Saturday, August 26, 2023 at 7:24 AM
To: Weka machine learning workbench list. <[email protected]>
Subject: [Wekalist] Re: Does Weka do assumption check?


> On Aug 24, 2023, at 6:28 PM, Shi, Catherine <[email protected]> wrote:
>
> Yes. Does Weka support these? Thanks!

I had a little trouble understanding the question. Any implementation of linear regression is going to make these assumptions. They are not going to check them. Normally it is up to you to decide if the assumptions are met and you want to use linear regression. Probably more normally you just try linear regression and see what results you get and don’t worry about the assumptions.

Pre-checking if the assumptions are met is sort of an interesting idea. I think Eibe was trying to show how you could do this with Weka.

There are many other classifiers that can be used so it usually easier just to try a few and see which perform best. Weka’s experimenter can be used for this. Or just repeated trial and error. You can also ensemble or use multiple classifiers together to improve performance.

If linear regression is a requirement and you want to be diligent in checking the assumptions. Eibe’s s suggestions would be the way to go.



_______________________________________________
Wekalist mailing list -- [email protected]
Send posts to [email protected]
To unsubscribe send an email to [email protected]
To subscribe, unsubscribe, etc., visit https://secure-web.cisco.com/1ZzU_JN1O26LT1Q7xvDaVjFIAQlhFXroZcHVLTVILtY_OP1Txf-GoIj2KMCg1Vzhc15McDjTw4SD2AOW4R8sX_qJwfz4vElSgFnji7DYaobX0zEe87vIiOTsjyqLz9rhHomSmEDVBg6_lWzg_uvC9kriWWNC4K1krT7nwr0z5ZIfcBLKBHClvyqflLJXGcwzzg4YsBZHYTKDDV_coshjucwqwh84f_FO06KaCQBYB5nAMGIr7-d7NGdAC1achgeUggxyWS1ZVIQobJmuYAq8za3lk2RIpgI7fHb8r9Q69e8ABB0Fjz8R1I1Vp7f6NxwZh/https%3A%2F%2Flist.waikato.ac.nz%2Fpostorius%2Flists%2Fwekalist.list.waikato.ac.nz
List etiquette: http://secure-web.cisco.com/1Hn63bgM9rRqljpvcNPWhc-vZ9MP4dbzjNSjOVO6wX-_RLdkw6rkUOaLhjrVfOHbrk3PhundHgoB3Xh1v7DbieZlAIdhqGRe7e4tG1mcq6SmSTkNaN-mYEmBiFFqIxfcvNM4IO6e1R_NZCjjlv7TB0KAZf33u7kYxMYqtLPYl697pMWWfBvj4joXh-3V9pITQkMRYpHG9xxw8u8kpj0cBttwpFP0X1UCOEp6UZm7gMN9qnoMpoFKM7s_fG_Wp9fu7_9n8GmPG08HcddKeDl0P1wsXJ8wL2qodHquQzSZaQLDqGSkIxVdL72MOMgvk39DJ/http%3A%2F%2Fwww.cs.waikato.ac.nz%2F%7Eml%2Fweka%2Fmailinglist_etiquette.html

_______________________________________________
Wekalist mailing list -- [email protected]
Send posts to [email protected]
To unsubscribe send an email to [email protected]
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html