'pointless' Need some help collecting narrowboat data

[David Mack]

On 02/12/2022 at 14:34, Mike Todd said:

 

Another of my pet hates are surveys that ask respondents to rate something on a scale, usually with no good reason on a 1-10 (or perhaps 0-10) but without any indication of what a particular level might mean. In some cases the results are even explicitly skewed - a car retailer tells its customers that they need to give it 10 out of 10 as anything less will be used by the head office to indicate less than satisfactory! Guess what most replies are?

Very few products or services are perfect. Anything that does the job well enough gets an 8 out of 10 from me (when I can be bothered to fill in the survey). But apparently that isn't enough!

[toppy007]

Hi everyone, firstly just wanted to say thank you too everyone who contributed.

 

I've decided to take a different direction on data collection because I'm not going to get the numbers required. If it was summer I'd have sat out at a lock. bit of a shame in terms of a useable product. But many ways to skin a rabbit. 

 

I've progressed in the sense of data collection. And have been using NLP and Fuzzy Parsing to obtain data. I'll now be cleansing the data. And selecting a training set from this. 

 

I've been drawn towards integrating a reinforced learning algorithm to help optimise the output.

 

Will update all on the results after Christmas only via the people added an email to the survey.

 

This is because of the stress I seemed to have caused many on the site.via my rubbish Gramma, spelling and subject matter. 

 

 

[Mike Todd]

1 hour ago, toppy007 said:

Hi everyone, firstly just wanted to say thank you too everyone who contributed.

 

I've decided to take a different direction on data collection because I'm not going to get the numbers required. If it was summer I'd have sat out at a lock. bit of a shame in terms of a useable product. But many ways to skin a rabbit. 

 

I've progressed in the sense of data collection. And have been using NLP and Fuzzy Parsing to obtain data. I'll now be cleansing the data. And selecting a training set from this. 

 

I've been drawn towards integrating a reinforced learning algorithm to help optimise the output.

 

Will update all on the results after Christmas only via the people added an email to the survey.

 

This is because of the stress I seemed to have caused many on the site.via my rubbish Gramma, spelling and subject matter. 

 

 

Be interested to know what steps you are taking to eliminate bias arising from the cleansing process and the subsequent selection. To test the methodology i your app this is not a problem but might it be so for a live app?

[toppy007]

Sorry todd 

 

If you've added your name to the email list it'll all be included in that but I'm not wasting my time with people on this forum who want to pick peoples Gramma apart for fun.

 

. 

[Jim Riley]

It will be the same picky pedantic Grumpfuttocks you meet going through locks! Boats attract them. I did warn you earlier🤣🤣🤣🤣

[MtB]

2 hours ago, toppy007 said:

Sorry todd 

 

If you've added your name to the email list it'll all be included in that but I'm not wasting my time with people on this forum who want to pick peoples Gramma apart for fun.

 

. 

 

 

I suspect your sloppy approach to grammar is matched by a similarly sloppy approach to data hence my reluctance to waste my time with your sloppily-crafted survey. Thus confirming Jim's post, one ahead of this!