User etang

17 votes

2 answers

2k views

Why not use evaluation metrics as the loss function?

asked Jun 2, 2022 at 21:13

5 votes

3 answers

658 views

How to choose between an overfit model and a non-overfit model?

asked Oct 31, 2020 at 5:07

4 votes

1 answer

715 views

Can log loss be an evaluation metric for classification models?

asked Jul 10, 2021 at 5:20

4 votes

1 answer

423 views

Which Method, Skip-gram or Negative-sampling, does Keras's Embedding Layer Use?

asked Oct 3, 2020 at 0:29

3 votes

1 answer

608 views

How to Deploy Contextual Bandits in Online Experimentation Platform?

asked May 22, 2020 at 0:19

3 votes

1 answer

155 views

How to test a hypothesis for mean difference within a range

asked Feb 13, 2022 at 5:55

3 votes

1 answer

3k views

Median absolute deviation only can be used for anomaly detection for time series without a trend?

asked Oct 28, 2020 at 19:04

2 votes

2 answers

932 views

What are the pros and cons of using a supervised learning approach for building a recommender system compared to traditional approaches?

asked Oct 29, 2021 at 17:16

2 votes

1 answer

302 views

How to determine decay rate?

asked Nov 6, 2019 at 22:10

2 votes

1 answer

50 views

Why Some Simulation Scenarios Don't Look Real? [closed]

asked Jul 18, 2020 at 19:16

2 votes

1 answer

210 views

Can I still use an overfitted model with high test accuracy?

asked Aug 29, 2020 at 17:07

2 votes

1 answer

289 views

Are Bandit Algorithms Considered as Online Algorithms?

asked Oct 19, 2020 at 19:42

2 votes

1 answer

44 views

How to find the closest table name for a given mis-spelled table name?

asked Nov 1, 2020 at 4:31

2 votes

0 answers

24 views

What metrics should I use when developing the first model?

asked Nov 15, 2020 at 5:15

2 votes

1 answer

44 views

How to Increase the weight for one predictor variable?

asked Nov 27, 2020 at 2:41

2 votes

1 answer

63 views

Need to do hyperparamter tuning for new features?

asked Feb 9, 2021 at 17:29

2 votes

1 answer

36 views

Online Tree Based Algorithms

asked Mar 7, 2021 at 6:14

1 vote

1 answer

62 views

Latent factors are the same in both decomposed matrices?

asked Jul 31, 2021 at 5:02

1 vote

1 answer

91 views

Why BERT Boosts Performance for NLP Tasks?

asked Oct 23, 2020 at 21:02

1 vote

0 answers

21 views

How to know if LDA(Latent Dirichlet Allocation) is applicable to a particular problem? [closed]

asked Oct 27, 2020 at 5:14

1 vote

0 answers

5k views

How to Tune Isolation Forest?

asked Oct 7, 2020 at 3:56

1 vote

1 answer

165 views

Speed Up KNN and Maintaining Accuracy for Anomaly Detection

asked Sep 29, 2020 at 15:39

1 vote

0 answers

1k views

Is Cook's Distance a reliable way to find influential points?

asked Oct 1, 2020 at 20:29

1 vote

1 answer

49 views

Will Different Types of Labels Affect Feature Engineering Outcomes?

asked Aug 12, 2020 at 17:15

1 vote

0 answers

30 views

Can Bayesian methods be used for fixed horizon hypothesis testing? [closed]

asked Nov 2, 2019 at 21:36

1 vote

1 answer

95 views

How to Retrain a Production Model?

asked Jun 14, 2020 at 4:08

1 vote

1 answer

569 views

How to Compare Variances?

asked Jun 26, 2020 at 20:47

1 vote

1 answer

153 views

Estimate remaining time to failure based only on average time to failure

asked Jan 10, 2020 at 0:34

1 vote

1 answer

401 views

What Does Oracle in Multi-Armed Bandit Literatures?

asked May 24, 2020 at 1:04

1 vote

1 answer

71 views

How to Create a Model Which Can Produce Accurate Predictions?

asked Jun 1, 2020 at 22:47

Stack Exchange Network

44 Questions

Why not use evaluation metrics as the loss function?

How to choose between an overfit model and a non-overfit model?

Can log loss be an evaluation metric for classification models?

Which Method, Skip-gram or Negative-sampling, does Keras's Embedding Layer Use?

How to Deploy Contextual Bandits in Online Experimentation Platform?

How to test a hypothesis for mean difference within a range

Median absolute deviation only can be used for anomaly detection for time series without a trend?

What are the pros and cons of using a supervised learning approach for building a recommender system compared to traditional approaches?

How to determine decay rate?

Why Some Simulation Scenarios Don't Look Real? [closed]

Can I still use an overfitted model with high test accuracy?

Are Bandit Algorithms Considered as Online Algorithms?

How to find the closest table name for a given mis-spelled table name?

What metrics should I use when developing the first model?

How to Increase the weight for one predictor variable?

Need to do hyperparamter tuning for new features?

Online Tree Based Algorithms

Latent factors are the same in both decomposed matrices?

Why BERT Boosts Performance for NLP Tasks?

How to know if LDA(Latent Dirichlet Allocation) is applicable to a particular problem? [closed]

How to Tune Isolation Forest?

Speed Up KNN and Maintaining Accuracy for Anomaly Detection

Is Cook's Distance a reliable way to find influential points?

Will Different Types of Labels Affect Feature Engineering Outcomes?

Can Bayesian methods be used for fixed horizon hypothesis testing? [closed]

How to Retrain a Production Model?

How to Compare Variances?

Estimate remaining time to failure based only on average time to failure

What Does Oracle in Multi-Armed Bandit Literatures?

How to Create a Model Which Can Produce Accurate Predictions?