Cross Validated
Sponsored by
Sponsored logo
Profile Activity
Meta user
Network profile
1,386
reputation
5
16

gazza89

Former physicist, slowly recovering as a data science consultant. First principles thinker over "throwing a neural network at the problem and hoping for the best"

46
answers
8
questions
~33k
people reached
  • London, UK
  • Member for 4 years, 8 months
  • 96 profile views
  • Last seen Oct 15 at 9:25

Communities (4)

Top network posts

We respect a laser-like focus on one topic.

Top tags (106)

Score 15
Posts 7
Posts % 13
Score 7
Posts 5
Score 7
Posts 4
Score 6
Posts 9
Score 5
Posts 5
Score 5
Posts 4
View all tags →

Top posts (54)

All Questions Answers
Votes Newest
24 Kullback-Leibler divergence WITHOUT information theory Nov 16 '16
12 Why does XGBoost have a learning rate? Jul 4 '18
6 Loss function and evaluation metric Nov 28 '18
5 Showing that ridge regression is a solution to the following optimization problem Sep 18 '18
5 Why is cross entropy not a common evaluation metric for model performance? Oct 11 '18
4 How is that likelihood is fixed by the model Oct 1 '19
3 100% accuracy on training, high accuracy on testing as well. What does this mean? Jul 23 '19
3 Is it important to make a feature scaling before using Gaussian Mixture Model? Oct 11 '18
3 If $A, B$ are conditionally independent by $C$ prove that $P(A|B,C) = P(A|C)$ Oct 5 '18
3 Does training loss go to zero in kernel regression? Aug 7
View all questions and answers →

Badges (21)

Gold

—

Silver

5

Rarest

Bronze

16

Rarest

View all badges →