Stands for 'Classification And Regression Trees'. CART is a technique for developing a tree model (T) to predict categories (C) and/or continuous values (R) by recursive partitioning. It does not make restrictive parametric assumptions. CART is a popular data mining technique.

learn more… | top users | synonyms (2)

0
votes
0answers
6 views

Tree classifier or nested model?

I am new to statistics and was wondering what the right kind of model to use for the following scenario. I have two sets of continuous observations A and B for 50 samples. These 50 samples are ...
0
votes
1answer
19 views

longitudinal dataset with multiple binary dependent variables

I have a dataset where a test (a continuous variable) is administered every 3 months for 2 years for most of the participants (approx 25% have one or two missing scores). There are further ...
1
vote
1answer
60 views

How to evaluate the goodness of fit for survial functions

I am a newcomer to survival analysis, although I have some knowledge in classification and regression. For regression, we have MSE and R square statistics. But how we can say that survival model A ...
0
votes
0answers
16 views

Fuzzy search in the database

Did anyone have experience with fuzzy search in database? Let's consider some type of Akinator. At each moment the idea is to choose the question which would split all possible (remaining) final ...
0
votes
0answers
8 views

How to control number of nodes which creating decision trees in r? [closed]

I am new to R. I am working on generating a decision tree from my dataset using any of the libraries like rpart, tree and party. However, my task is to limit the number of nodes in the decision tree ...
0
votes
2answers
33 views

Analyses of mixed variables

I'm relatively new to statistics, and am currently working some data collected as a part of an interview survey. I have a response variable in ordinal form, which mostly looks into people's ...
2
votes
0answers
13 views

Continuous variable evaluation in decision trees

I was going through the C4.5 and ID3 algorithms used to construct a decision tree. Was wondering if there is an efficient way to compute information gain from a continuous variable (during the step ...
1
vote
0answers
4 views

Dependent predictors with converse effects on the target

I am trying to create a predictive model for marketing in the natural gas field. The model is supposed to guess how probable it is to make a contract in that particular building given many internal ...
0
votes
0answers
10 views

Using all the data for a decision tree prediction

When doing regression analysis with training and test sets, I would fit the model to the training data. When happy with the model fit I'd then run it against the test/holdout set to get a true ...
0
votes
0answers
15 views

What statistical test(s) should I run to analyse a directed-tree like data?

Suppose I have data of a kind of binary tree format. There are three levels in the three, thus four different leaf nodes. I would like to find the optimal path from the root node to one of the leaf ...
0
votes
0answers
8 views

VC-Dimension of n-node binary decision tree in N-dimension feature space

Given input feature space $\mathcal{X} =\{0, 1\}^N$ and output label space $\mathcal{Y}=\{0,1\}$ , prove that the VC-dimension of a binary decision tree with $n$ nodes is in $O(n\text{log}N)$. I've ...
2
votes
0answers
52 views

How to split a decision tree when information gains of all attributes are zero?

The textbook tells us that we should choose an attribute with the maximum information gain to split a decision tree. My question is what if all information gains are zero? Should we stop splitting or ...
0
votes
0answers
30 views

Query on Quants

I am working on decision trees for the first time at job. I have done lot of research on CHAID and CART algorithms but find different answers to a very simple question given below : What kind of ...
2
votes
0answers
18 views

mob model tree algorithm

I am trying to figure out the inner workings of the mob function in the party package. I can't figure out how the splitting variable is selected when it is a categorical variable. In the publications ...
0
votes
0answers
86 views

Gini index vs entropy

If I have a discrete probability distribution $p$ with $K$ classes Gini index = $\sum_{K}$$p_k$(1-$p_k$) Entropy = -$\sum_{K}$$p_k$log$p_k$ Per 'The Elements of Statistical Learning', ...
0
votes
0answers
5 views

number of nodes in an unpruned decision tree

What is the number of nodes in an unpruned decision tree that is trained using n samples and that grows until there is only one sample in each leaf? I would like to know if there is a formula to ...
0
votes
0answers
32 views

Sample Weights for classification using Gradient-Boosted trees?

How can "weights" be given to different samples according to their relative importance while using Gradient boosted decision trees for classification? How does the ...
4
votes
1answer
82 views

Are Auto-Associative Regression Trees Distinct from Auto-Regressive Trees?

After some reading in the field I was confused as to whether these two models are distinct or really the same. I'm just looking for a simple yes/no with a brief explanation. Note that ...
0
votes
0answers
32 views

How are CP (Cost Complexity) values calculated in RPART (or decision trees in general)

From what I understand, the cp argument to the rpart function helps pre-prune the tree in the same way as the minsplit or minbucket arguments. What I don't ...
0
votes
0answers
27 views

feature selection for longitudinal data

I have a longitudinal data which looks like this. Number of time points are different for each ID. Y is the binary response variable (take values 0 & 1) and X1-X20 are either continuous or ...
1
vote
1answer
26 views

Obtaining easy-to-interpret decision tree

I am trying to create a decision-tree out of a number of attributes, where there are only two final classes and the classes are highly unbalanced (Class 1: 95.5%; Class 2: 4.5%). The idea is to ...
1
vote
1answer
47 views

Post hoc selection of important features in random forest?

I want to guarantee a parsimonious random forest (few features used). What are methods to do this? It was suggested to me to get the feature importance after the model was created, and then create a ...
0
votes
0answers
19 views

Considering non-i.i.d. covariates in random forests

Random forests are theoretically funded on the assumption that the data are i.i.d. realizations from a multivariate random vector $(X_1, \ldots, X_p, Y)$. Does it make sense to use random forests (for ...
1
vote
0answers
33 views

categorical feature ranking

I would like to rank categorical features by the order or importance in a classification/regression setting. Input There are two features, which are survey questions: "how is your mood?": four ...
0
votes
0answers
14 views

Minimum population for decision trees and association rules

Hi I'm quite new to this and I'm playing around with R and Microsoft's SSAS. Does anyone have a rule of thumb how big a data set has to be for association rules and decision trees to be statistically ...
0
votes
1answer
29 views

A question on no. of training examples and decision trees

I have a set of around 200,000 training instances. Each training instance consists of an attribute called $duration$, a discrete integer type and a time series of floating-point values, in form of a ...
0
votes
0answers
19 views

On population variable importance

Consider we run a random forest on $n$ independent realizations of a random vector $(X_1,X_2,X_3,Y)$ assuming $Y$ is a numerical response variable. Let $f$ be the best theoretical classifier defined ...
0
votes
0answers
13 views

How does RPART pick a splitter when there are at least 2 splits having maximal information gain?

I'm not sure the best way to explain this, so let me give an example that motivates my question. I have tried reading the RPART manual, documentation, and its code, but I have not been able to resolve ...
0
votes
0answers
5 views

Design a feature with time and presence information

Context: I am working on a decision tree classifier, trying to classify businesses as to whether they are likely to have an event occur (default) in the next 90 days. One input I get is whether, and ...
0
votes
0answers
23 views

Evaluating mean-squared error

Hello I am running a Regression Tree experiment. I am new to Regression Trees, and I am using Mean Squared error to test my tree. I am confused because I am getting a large Mean Squared Error but I ...
2
votes
1answer
101 views

Why should all Cross-Validation results be higher than the result on the test dataset?

Sorry, I'm not an expert and my question could be fundamentally wrong. I've read this interesting question because I also was wondering whether to train the model again after cross-validation. Now, ...
1
vote
1answer
38 views

Is there such thing as correlation trees? Clustering rows of X based on correlation between A and B

I have been searching for several days for a method that fits this description, though cannot find one. I'm pretty sure it must exist. The problem (short version): I'd like to run something like a ...
0
votes
1answer
21 views

Decision Trees on training data

Wouldn't any decision tree trained on a training data set have no errors in classification? In other words, wouldn't every data point be classified correctly in the training data set? How would this ...
0
votes
0answers
26 views

Misclassification Rate

The misclassfication rate in decision trees is defined as $$1-\max_{j} p_{j}$$ Suppose we want to classify people into republicans and democrats. In the training data, $\hat{p}_1 = 0.7$ (probability ...
1
vote
0answers
8 views

Comparing and contrasting classification trees over the years

At work I have decided to use a classification tree to do some analysis on the people at the bottom decile for hourly pay. Mainly for inferential reasons. My datasets are broken up in years, I have ...
1
vote
0answers
31 views

In CHAID, shouldn't we merge categories when p<alpha rather than p>alpha?

In CHAID, the categories are merged when P>alpha in the first step. BUT Since CHAID uses Chi-square statisitic, if p-value < alpha, we reject the null ( Independence), hence, meaning the ...
1
vote
1answer
24 views

What methods can be used to transform data?

I am solving a binary classification problem with 4 predictor variables. The variables didn't seem to be linearly separable. I have used Neural Networks and Kernel SVM which work and give desired ...
0
votes
0answers
25 views

Down-sampling with building models (specifically random forests)

I was wondering if anyone had ever used down-sampling to build random forests with data that has unbalanced classes. Basically down-sampling samples (with replacement) x*min from the population where ...
0
votes
0answers
27 views

building a decision tree after association rules

I built association rules using R Arules package. Then I filtered those rules and kept rules that have a specific variable on RHS. That variable is my Y variable. In entire dataset, my variable occurs ...
0
votes
0answers
24 views

Improving sentence segmentation in NLTK

I have been looking into problem of sentence segmentation lately. I have been referring to NLTK's book for this purpose. I followed their procedure to segment sentences presented here: ...
3
votes
0answers
73 views

Why will a random forest not outperform a regression tree?

I have a training dataset with a binary response variable, 6 independent variables, and 21,000 observations. I've fit both an ordinary regression tree and a random forest (mtry = 2, ntree = 2000) and ...
0
votes
0answers
71 views

Decision tree in R

I am new to machine learning in R. This is my data set. ...
2
votes
2answers
136 views

How to prune a decision tree properly in R

I have a sample of 12,500 observations and 12 explanatory variables. I want to build a pruning decision tree, to do that I am using the rpart function and then the ...
0
votes
0answers
14 views

decision tree : categorical vs numeric explanatory variables [duplicate]

I found similar topic but it dealt with the question whether it's better to discretize numerical dependent variable or leave it as it is. I'd like to find out what are your experiences with training ...
1
vote
1answer
57 views

How to promote a regression tree over a GLM?

Does anybody have any suggestions about promoting the use of a regression tree over a GLM when the two models fit the data almost exactly the same? My team's current arguments are a) a tree is ...
0
votes
0answers
17 views

Heuristic for bagging in random forests

Is there a good heuristic for what fraction of the data to use when training a random forest with bagging? I imagine this fraction should depend on the number of trees in the forest, like 2/3 ...
1
vote
0answers
54 views

How do i estimate the Weights of the predictions assigned to each of the tree in GBM using R? How does GBM split nodes?

I ran a GBM model in R with loss function as bernoulli and n.trees=1000. I want to see the weights assigned to the predictions coming from 1000 trees. Is there any command in R that does that? How ...
0
votes
1answer
28 views

multiple predictors in decision tree model

I am using like 10 predictors in my decision tree, but the rpart function uses only like 8 of them. does it mean that rest 2 are not needed or they become redundant? like i have age and child coded ...
1
vote
1answer
66 views

Difference between rel error and xerror in rpart regression trees

Wondering what the difference is between rel error (relative error) and xerror (apparent error) in regression trees? I am using the rpart package and the output returns these metrics ...
0
votes
0answers
18 views

CART & stationarity assumptions

I have a question regarding if there are any stationarity requirements for a regression tree. If I were to regress 2 share prices on each other (normal linear regression), it would be classified as ...