Skip to main content

Questions tagged [h2o]

H2O is an in-memory platform for distributed, scalable machine learning.

Filter by
Sorted by
Tagged with
0 votes
0 answers
14 views

DRF all features have positive contribution to SHAP values except one

I have trained a DRF model using h2o and the shap summary plot came out like this. Why aren't feature contributions centered? It seems weird that one feature contributes negatively and all the others ...
Pedro Schuller's user avatar
1 vote
1 answer
37 views

Predicted class probability in threshold moving

I am training a model for the task of Binary classification using H2O.ai. The final output to the user is the probability of class_1. Recently, I found that by ...
Muhammad Ahsan's user avatar
1 vote
0 answers
80 views

Training Regression Models to Predict Continuous Probability Values in [0, 1] [closed]

I'm working on a machine learning project where my target variable represents continuous probability values that must fall within $[0, 1]$. While I understand regression models are suitable for this ...
Anastasiya-Romanova 秀's user avatar
1 vote
1 answer
90 views

Identifying root cause of very poor Random Forest model

I've been working on building a random forest model using h2o.ai in R for climate data. I know that there is some issue, either with my understanding of randomforest, code or dataset. However, I'm not ...
DGeospatial's user avatar
0 votes
0 answers
70 views

Finding the Variance-Covariance matrix of IRLS output

How can I obtain the Variance-Covariance matrix of estimated coefficients, $\hat{\beta}$, which are a result of some GLM (solved for example using IRLS. Is there a reference you can direct me to? In R ...
Kozolovska's user avatar
  • 1,355
0 votes
0 answers
123 views

H2O Deviance - Negative Binomial

I'm hoping to get some clarifications on the deviance calculation of negative binomial. From H2O documentation, the deviance formula for negative binomial regression is expressed as: $$D=2\sum_{i=1}^{...
Naomi's user avatar
  • 1
0 votes
1 answer
114 views

Understanding the idx column in h2o metrics [closed]

h2o model metrics results report generated like this. ...
John's user avatar
  • 3
1 vote
1 answer
112 views

Understand abrupt jumps at the end of the score history of h2o deeplearning models

I train deep learning models on the fashion MNIST dataset using H2O with the R interface. My score history often looks strange: the end is jumping up or down as two different values are reported for ...
janosdivenyi's user avatar
2 votes
1 answer
264 views

Multi-level / Hierarchical Machine Learning

I am trying to tackle a problem that involves binary classification. However, my data is multi-level or hierarchical in its structure. This example illustrates its structure: ...
DaveTheRave's user avatar
2 votes
0 answers
371 views

Model Stacking and tuning a meta-model - CV strategy?

I was hoping some of the more experienced ensemblers could help me with a couple of questions I have regarding stacking. The assumption is that we have a classic ...
Liam Morgan's user avatar
1 vote
2 answers
584 views

Which loss functions does h2o.gbm use by default?

The GBM implementation of the h2o package only allows the user to specify a loss function via the distribution argument, which defaults to ...
user313258's user avatar
2 votes
1 answer
191 views

H2O GLM for heavy-tailed data [closed]

I am trying to run H2O GLM (OLS, lasso, ridge, EN) for stock returns, which have very heavy tails (i.e. potentially infinite variance). Is there a robust loss function modification for this, say Huber ...
HeavyTailedH2O's user avatar
0 votes
0 answers
592 views

How H2O perform class balancing?

I wanto to perform class balancing using h2o autoML. I know there is a parameter class_sampling_factors that allow to specify the under/over sampling factor for ...
A1010's user avatar
  • 213
3 votes
1 answer
846 views

How should I decide the decision-threshold in a classification built using stacked ensemble classifier

I am attempting a stacked ensemble model to achieve binary classification for the first time. Should the decision threshold be One for which I receive max F1 score (assuming F1 score drives the ...
learner's user avatar
  • 617
1 vote
1 answer
420 views

H2o Python - How do I determine the threshold for AUC

I am new to H2o and I having trouble with AUC and Confusion Matrix I have a classification model using H2o in Python for which the AUC = 71% But the accuracy based on confusion Matrix is only 61%. I ...
TigSh's user avatar
  • 113
2 votes
1 answer
403 views

Why fit new TargetEncoder for test data?

In the H2O tutorial on target encoding they recommend fitting a new TargetEncoder to the entire training set to encode the test data. Why not just use the averaged TargetEncoder fit to the training ...
andrew's user avatar
  • 163
0 votes
1 answer
61 views

How to determine drivers of h2o random forest predictions?

I have an h2o random forest algorithm. I trained the algorithm, tested it, and interpreted the performance. The algorithm is a binary classifier, so it's spitting out 1s and 0s for each record in the ...
OverflowingTheGlass's user avatar
0 votes
0 answers
377 views

Predicted probabilities very close to 0 and 1 in GLM model

I've added new attributes to the binary GLM model. AUC climbed to 98%, logistic loss decreased to 0.45. Training set has ~50 cases. I can see that predicted probabilities are extremely close to 0 and ...
MLearner's user avatar
1 vote
0 answers
15 views

Strange distribution of prediction probilities

I am using H2O AutoML (python 3.28.0.2) to build a bernoulli GBM model for a dataset. The dataset is relatively small train/validate/test: 6217 1310 1550 My settings are automl_params = { '...
Robin Aly's user avatar
1 vote
1 answer
830 views

Default threshold in cross-validation metrics - h2o R package

I created an cartesian grid of GBMs using h2o package in R and saved cross-validation metrics for each model in a data frame. So, for each model, I stored the ...
user260147's user avatar
3 votes
1 answer
4k views

Variable Importance Intepretation for GLM

I'm having a confusion and can't find any answer from the docs.ai that H2O Team provided. I'm creating a summary from my glm and receive a variable importance table. But i can't understand what is ...
Yohanes Lim's user avatar
2 votes
1 answer
247 views

Neutralise/remove feature from GBM

I need to remove a feature (variable) from my GBM without rebuilding the model and excluding the variable, what would be the best approach to do this?
Charl Francois Marais's user avatar
2 votes
0 answers
163 views

variable importance -frequency and variable importance -cover (h2o)

For some models h2o can produce three forms of variable importance Analysis ( variable importance , variable importance -frequency and variable importance -cover) , can anyone point me to a reference ...
Robert Treloar's user avatar
0 votes
1 answer
180 views

Inconsistency between poisson and negativebinomial in glm

I am working with the negativebinomial distribution for GLM. I have done one test which is finding the poisson distribution results. Here is the first test: ...
Quantic's user avatar
0 votes
1 answer
459 views

Improving Average F1 Score for Multiclass Classification

I'm trying to do a multiclass classification with h2o in R. I stacked a model with a RF, GBM and deeplearning. The accuracy is ok (~0.81), but the average F1 score is bad because class B has a very ...
abigfatcat's user avatar
3 votes
3 answers
1k views

LASSO or random forest (RF) to use for variable selection when having highly correlated features in a relatively small dataset with many features?

I have a cross sectional data-set with around 1000 features and 5000 observations. There are many features (no categorical features) which are highly correlated (higher than 0.85). I want to decrease ...
mlee_jordan's user avatar
0 votes
1 answer
1k views

h2o glm tweedie for categorical variables?

To build a tweedie glm for categorical variables, the document suggested that I can use data['variable_name'].asfactor(). However, in the model output, there is ...
Jie Huang's user avatar
0 votes
1 answer
394 views

How does H2o handles missing values in DRF? [closed]

Just wanted to confirm that the h2o's implementation of RF (DRF) handles the missing values for both categorical and numerical features the same i.e., as a separate category?
TaraM's user avatar
  • 1
2 votes
1 answer
251 views

Why are AUC and logloss metrics not available in the "maximum metrics" table produced by H2O? [closed]

I am running the h2o.gbm algorithm using five-fold cross validation to predict a binary outcome. I want to see what threshold to use as a cutoff for classifying predictions, and I am wondering why the ...
NewToGBM's user avatar
0 votes
1 answer
653 views

How are the training and cross-validation metrics calculated in H2O?

I am working with the GBM algorithm in H2O in R. I am using 100% of the data as the training data, and then using 5-fold cross-validation to train and validate my model using 100% of the data. My ...
NewToGBM's user avatar
0 votes
0 answers
1k views

Feature selection in xgboost vs GBM in H2O

I am working on a big data set( more than 100 variables) and 30 million observations. I tried to build 100 models with a grid search using both XGBoost and GBM in H2O (Sparkling Water). I realized ...
Rio's user avatar
  • 141
3 votes
1 answer
3k views

New factor levels in testing data set not present in training data in h20.randomforest

In randomforest classification using h20 package, there are factor levels which are present in testing data but not in training data.There is a warning message in predicting the values of testing data,...
PA17's user avatar
  • 31
3 votes
1 answer
552 views

binomial responses in h2o gbm

I am modeling the probability of success in a dataset where I have a both the number of trials and the number of successes (and, obviously, I am modeling $p_i=\frac{\textrm{total successes}}{\textrm{...
Giorgio Spedicato's user avatar
1 vote
1 answer
105 views

What kind of impact do autoencoders have on final model performance when compared to models trained only on supervised data? [closed]

For example, say we have two datasets, a labeled set (I will call it df_labeled) of nrows=200k and an unlabeled dataset (df_unlabeled) of nrows=800k and we want to build a binary classifier. I clearly,...
Sourav's user avatar
  • 13
0 votes
1 answer
194 views

Quadratic Weighted Kappa metric in H2O package for model performance

I am running a multiclassification problem and before I make a function by myself I was wondering if anyone knows of a pre built quadratic weighted kappa function in the h2o library.
Molia's user avatar
  • 101
1 vote
1 answer
495 views

H2o k-fold validation [closed]

I need to get some clarification on how H2o creates a training model from the k-fold validations. Below is my understanding, please correct where I am wrong: If I set nfolds = 5, then H2o will split ...
MD_1977's user avatar
  • 11
2 votes
1 answer
165 views

DeepLearning & Anomaly Detection - Understanding & How to Properly Tune

I'm looking into understanding the Deeplearning anomaly detection algorithm provided by h2o. I tried to recreate an example below. Perhaps some of these questions are basic, but I'm trying to better ...
user3788557's user avatar
  • 1,629
1 vote
1 answer
938 views

H2O GBM and Caret GBM

Hi I have doubt regarding the interaction. depth parameter in caret. I found a useful link hereabout interaction.depth in caret Now I am trying to find the similar parameter in H2O-GBM . Can anyone ...
Kitooos's user avatar
  • 11
3 votes
1 answer
3k views

categorical_encoding in h2o - what is the difference between the options

I'm trying to understand the pros/cons and when to use the various encoding options that are available to me in h2o with the parameter 'categorical_encoding'. It would be helpful if people could ...
user3788557's user avatar
  • 1,629
2 votes
2 answers
902 views

May somebody help with interpretation of trees from h2o.gbm, see as photo attached

This picture is from h2o.gbm, while I'm not sure how to interpret the numbers in it. What is the big title "Class NO" mean? Does it mean the root node is labeled "No"? Or does it mean this tree is ...
EmLp's user avatar
  • 55
-1 votes
1 answer
511 views

Issues with XGBoost on H2O environment

I have a dataset from which I built lags at different levels to use as features in the XGBoost model. When I ran XGBoost models on H2O, the model is picking up the features which contain higher values ...
Praveen Surendra's user avatar
1 vote
1 answer
145 views

Is setting a self learning system possible via incremental (online) learning?

Self learning and incremental learning are all new to me. I am trying to develop a system for one of my case. Simply I have a data set (with about 90K observations and 400 features) for a binary ...
mlee_jordan's user avatar
1 vote
2 answers
706 views

How NULLs in numerical variables are treated in tree-based models?

I understand that in tree-based models (CART, Gradient boosted trees, etc.), NULLs (i.e., NaN) in categorical variables can be treated as a separated category, while making node splits. However, how ...
ccy's user avatar
  • 138
1 vote
2 answers
175 views

h2o : Any components not running or stored on local machine? [closed]

The descriptive language used when referencing how h2o works is a little confusing to me at times (e.g., client, cluster, "internet"). I try to remain vigilant in not uploading or exposing my datasets ...
hlsmith's user avatar
  • 156
1 vote
1 answer
60 views

Help about cosine simulating in h2o [closed]

So I'm beginning in deep learning and especially in h2o. I tried to simulate cosine function in R, not to compute it like for example by using h2o.cos(), But after many and many more combinations of ...
Boris's user avatar
  • 11
6 votes
1 answer
247 views

Special values in continuous numerical variables/features in Random Forest

I have a binary response variable I am seeking to predict using Random Forest. I have a sizable dataset of 150k rows, I have about 200 independent variables or features to use to model the outcome. ...
JPErwin's user avatar
  • 463
1 vote
1 answer
3k views

Extracting H2o Cross Validation Results

I am using H2o library in R and have a slight confusion that you learned people might be able to help with. I am not sure how to interpret the output from ...
Nick RT's user avatar
  • 11
3 votes
1 answer
689 views

H2O PCA number of components

I wonder why number of components in H2o PCA algorithm is limited to 9. It is not sure sometimes to be enough. k: Specify the rank of matrix approximation. This can be a value from 1 to 9 and ...
NiMa's user avatar
  • 31
0 votes
0 answers
555 views

How to improve GBM performance

I'm trying to model insurance losses, with a Tweedie distribution. I have a data set of about 40 million records, and over 100 independent variables. My response variable is "loss", I take the log of ...
Will.I.am's user avatar
2 votes
1 answer
1k views

H2o interpretability - LIME

I have trained a model to predict heart attacks using random forest algorithm using H2O. I have good performance in cross validation. Now, I want to give more interpretation to the predictions in a ...
Jasam's user avatar
  • 21