Performance of Algorithm...

[AkshayChoulwar]

I am thinking that for classifying ingredients....

first we make set of diary items and whenever any items come we check the set and if value positive then we add it in category (if this is a requirement then why we need ML and NLP here) without ML also solution for this problem is possible

[atherdon]

there no innovation in simple programming and just classifying - is only a first step on the road first step in order to see if this can be useful

Just duplicating in order to have that here.

btw, maybe you right - so i don't know and maybe you don't know too - i mean we cannot find it out without trying :)

[atherdon]

but if you tell me - let's find another way - we'll sit down and find it

[AkshayChoulwar]

I implemented it using machine learning but still have to make modifications......

there are two advantages using Machine Leaning:

1.Once the model trained on testing it will take less time if our data is like(in GB) 2.If the user entered wrong spelling like for Butter and Buttered it will classify perfectly to dairy for both.

[atherdon]

So, what will be our next steps? you asked about sharing with you more ingredients in order to test how your algorytm works and understand will it find all dairy related, right?

about option 2 - perfect! will check my other repos and find an advanced list of ingredients

On Tue, Feb 6, 2018 at 7:29 AM, Akshay Choulwar notifications@github.com wrote:

I implemented it using machine learning but still have to make modifications......

there are two advantages using Machine Leaning:

1.Once the model trained on testing it will take less time if our data is like(in GB) 2.If the user entered wrong spelling like for Butter and Buttered it will classify perfectly to dairy for both.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/atherdon/gs-ml-nlp/issues/12#issuecomment-363315364, or mute the thread https://github.com/notifications/unsubscribe-auth/ABZrDtmSw48AImR_rZ9NV2JCESLyBjpWks5tR-NPgaJpZM4R2-Cc .

[atherdon]

did i get you right - i need to convert this data, right : @atherdon/gs-ml-nlp#9?

[AkshayChoulwar]

Yes...convert it and update the ingredients. csv (add other category data also)..we have only 110 ingredients in csv file so far.

[atherdon]

ok, so i got it right.

On Sat, Feb 17, 2018 at 5:20 AM, Akshay Choulwar notifications@github.com wrote:

Yes...convert it and update the ingredients. csv (add other category data also)..we have only 110 ingredients in csv file so far.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/atherdon/gs-ml-nlp/issues/12#issuecomment-366412277, or mute the thread https://github.com/notifications/unsubscribe-auth/ABZrDnhCvbrtavKD0LaTyXqaYd5VMaz-ks5tVkV4gaJpZM4R2-Cc .

[atherdon]

question: if i attach csv without category - will it be ok? i mean if i place category 'undefinded'

On Sat, Feb 17, 2018 at 7:57 AM, Arthur Tkachenko < arthur.tkachenko.netweight@gmail.com> wrote:

ok, so i got it right.

On Sat, Feb 17, 2018 at 5:20 AM, Akshay Choulwar <notifications@github.com

wrote:

Yes...convert it and update the ingredients. csv (add other category data also)..we have only 110 ingredients in csv file so far.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/atherdon/gs-ml-nlp/issues/12#issuecomment-366412277, or mute the thread https://github.com/notifications/unsubscribe-auth/ABZrDnhCvbrtavKD0LaTyXqaYd5VMaz-ks5tVkV4gaJpZM4R2-Cc .

[AkshayChoulwar]

Nope....we required csv for training and it should have label(category) so that we can train it.

for testing phase it will be OK. On Feb 17, 2018 11:36 AM, "Arthur Tkachenko" notifications@github.com wrote:

question: if i attach csv without category - will it be ok? i mean if i place category 'undefinded'

On Sat, Feb 17, 2018 at 7:57 AM, Arthur Tkachenko < arthur.tkachenko.netweight@gmail.com> wrote:

ok, so i got it right.

On Sat, Feb 17, 2018 at 5:20 AM, Akshay Choulwar < notifications@github.com

wrote:

Yes...convert it and update the ingredients. csv (add other category data also)..we have only 110 ingredients in csv file so far.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub <https://github.com/atherdon/gs-ml-nlp/issues/12#issuecomment-366412277 , or mute the thread https://github.com/notifications/unsubscribe-auth/ ABZrDnhCvbrtavKD0LaTyXqaYd5VMaz-ks5tVkV4gaJpZM4R2-Cc .

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/atherdon/gs-ml-nlp/issues/12#issuecomment-366420033, or mute the thread https://github.com/notifications/unsubscribe-auth/AQy79fKqlrT7RBy8xpbDaw6QGgd_bedeks5tVmxegaJpZM4R2-Cc .

[atherdon]

ohh, right now i understand what you need exactly from me. the easiest way for me is to export data from this link: https://airtable.com/shrmtzBGvND2be9GF/tbltaXhTFOl8FCdzV/viwscovAeDhZnk2Ae and add dairy as category, cause there only dairy items now. will it be ok as fast turn around?

[atherdon]

And i start realize that we need more data that i was prepared before...

[AkshayChoulwar]

Yes.. On Feb 17, 2018 12:09 PM, "Arthur Tkachenko" notifications@github.com wrote:

ohh, right now i understand what you need exactly from me. the easiest way for me is to export data from this link: https://airtable.com/shrmtzBGvND2be9GF/tbltaXhTFOl8FCdzV/viwscovAeDhZnk2Ae and add dairy as category, cause there only dairy items now. will it be ok as fast turn around?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/atherdon/gs-ml-nlp/issues/12#issuecomment-366421568, or mute the thread https://github.com/notifications/unsubscribe-auth/AQy79eOJvWl_RrfVWdtJ-iEhUZsq2-xaks5tVnQNgaJpZM4R2-Cc .

[atherdon]

i think it can be closed