Comments (3)
I see it closed, but for the benefit of others, maybe better explain it here.
Short answer: as explained in ariel.csv nosleep
means less than normal (in this example, 8 hours) of sleep, whereas sleep
means more than normal.
Generally, when doing ML with vowpal wabbit, weights start at zero and end-up relative to 0. The label is weight-gain (positive number) vs the previous data example. Factors that drive weight-gain are positively correlated with that label, and factors that drive weight-loss are negatively correlated with the label. If I use only one feature sleep:[hours_of_sleep]
it'll be end-up biased (centered at 8 hours) which may make the result very misleading. Worst case: if you suffer from chronic sleep shortage, and you gain weight, it'll appear as if you gain weight by sleeping.
Ideally, this "model vs the norm" principle should be applied to all features, I didn't do it that way because at the time I didn't fully grasp the importance of it, and because what I ate was too diverse to apply the principle to.
from weight-loss.
Arielf, It's so kind to receive your reply for a closed issue. That makes me rethink it which I thought I had understood.
Now I appreciate it more than before. I have known every factors should be "relative".
So a better model maybe like that:
1、Record everything with absolute quantity, such as:
R1: sleep:6(hours), bread:2(slices)
R2: sleep:7(hours), bread:3(slices)
2、Average it and rewrite it with relative quantity, such as:
R1: sleep:-0.5, bread:-0.5
R2: sleep:0.5, bread:0.5
3、Analysis
What do you think?
from weight-loss.
Yes. Thanks. I was thinking about it myself, and I think this is a better approach.
When I get time I plan to change the pre-processing code so it does average (or median) every factor globally first and then assign relative weights depending on their magnitude vs this (median/average) norm.
from weight-loss.
Related Issues (20)
- my fitness pal as data store HOT 1
- Extending the fasting time vs. eating fewer calories HOT 1
- confidence levels HOT 4
- Deriving importance factors on a per-week basis HOT 3
- Request for details about the specific algorithm used HOT 4
- How-To Section HOT 2
- OSError: No such file or directory HOT 9
- Age constant HOT 1
- Mac ggplot2 issue HOT 3
- semi-colon in Python HOT 1
- Neat analysis! How many days worth of data did you use? HOT 2
- Errors with Python HOT 2
- Please update your weight loss progress since July/Aug 2016
- RE: Please Update with most recent data HOT 1
- What does a unexpected "Constant" row mean in output file "scores.txt"? HOT 3
- How to do combinations? HOT 6
- Making sense of the results
- Numerical-Weight HOT 1
- Understanding the RelScore better
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from weight-loss.