0
votes
0answers
5 views

Train softmax temperature during function optimisation

Instead of simply choosing a softmax temperature, does anyone include this parameter during function optimisation? e.g. in the last layer of a neural network during backprop? Does it make sense to ...
1
vote
0answers
12 views

Three-level partially nested model

I am modeling change over time in group psychotherapy subjects using R and lme4. My data have the following structure: subject (id) time (code 1-10 for equally spaced repeated measures) outcome (for ...
0
votes
1answer
28 views

Logistic Regression Probabilities

I have performed a logistic regression on some data, and the function is: $p(x,y) = \frac{\exp(4.5 + 3.5x - 0.1y)}{1+\exp(4.5 + 3.5x - 0.1y)}$ This also gives you the probability of observing some ...
0
votes
0answers
4 views

What is the difference between R2 and variance score in Scikit-learn?

I was reading about regression metrics in the python scikit-learn manual and even though each one of them has its own formula, I cannot tell intuitively what is the difference between R2 and variance ...
-1
votes
0answers
11 views

What is 'E' embedding matrix in “Show, Attend and Tell” paper? [on hold]

I'm reading this paper Show, Attend and Tell: Neural Image Caption Generation with Visual Attention I couldn't understand what exactly the embedding matrix E is and how is it learnt ? Also what are ...
0
votes
1answer
27 views

true negative is 0% whereas true positive is 100% correctly classified

I used Naive Bayes from Spark's MlLib to train a model and test it on the data (in the form of an RDD). The results were confusing. the data and results are as follows: The problem is a binary ...
0
votes
1answer
11 views

Convolutional Neural Network for 3D point cloud?

Can Convolutional Neural Networks or Deep Architectures be used for generating 3D point clouds ?
0
votes
1answer
10 views

Mahalanobis Distance and feature scaling

I've been using Mahalanobis distance to look for outliers. This link: https://www.cs.princeton.edu/courses/archive/fall08/cos436/Duda/PR_Mahal/M_metric.htm says that feature scaling is addressed in ...
0
votes
0answers
16 views

Interpreting interaction term that only includes some of the variables in the model

I have a multinomial regression model with 5 variables (forced to enter the model), and all 2 way interactions between those variables (forward stepwise). The variables are: age, education level, ...
0
votes
0answers
2 views

Repeated measures with categorical variables in SPSS

I'm looking for some help with my thesis. The research is the following: I asked my respondents to give an alibi for several time periods; a day ago, a week, a month, 6 months and a year ago. For ...
0
votes
0answers
6 views

Cronbach Alpha Assumptions

I'm currently learning ho to use Cronbach Alpha in R. I have a scale with 7 items and 63 respondents. The aim is just to get a practical understanding of what Cronbach Alpha is. There are some nulls ...
0
votes
0answers
8 views

Create a predictive user model

I am a bit lost with creating a user model in R. I would like to create a model that predicts whether a user is likely to do an action or not, based on data on his past behaviour (Target variable ...
0
votes
0answers
9 views

Repeating games and R value [migrated]

I'm new fish in the water of Game Theory and just got stuck with calculating discounting rage (or discounting parameter) with 2x2 matrix. The main condition is that the game is repetitive. Here is ...
0
votes
1answer
16 views

Does the p value for logistic regression depends on odds ratio or logit?

The p value in a logistic regression is given for the B estimates. But what if I'm reporting odds ratio instead of logit. Do the same p values apply?
0
votes
1answer
16 views

Statistical analysis for replication in very small experimental datasets

A colleague does replication on a quite coslty experiments. There are four different conditions, each one duplicated. The outcome with $4\times 2 = 8$ points is illustrated below: The analysis is ...
0
votes
0answers
11 views

Forecasting: Exponential Smoothing overfitting

I have a 7 week time series(49 days) and I am told to forecast the next 2 weeks. One approach of mine is to use ets function (of package...
0
votes
0answers
9 views

What is the method to find difference in mean of test and control population where data were collected for a marketing campaign?

What is the method to find difference in the response rate of test and control populations in SAS where data were collected for a marketing campaign? (10-20% were control, and the rest were the test ...
0
votes
1answer
17 views

Are PAM silhouette and K-means silhouette related ? [on hold]

I used PAM Silhouette with different seeds to find the best k for ...
0
votes
1answer
17 views

Finding outliers in multiple dimensions

I'm working on dataset which isn't normally distributed. It contains three dimensions: cost, discount and profit. I'm trying to find outliers in all these dimensions. I used $\text{z-score}$ to find ...
0
votes
0answers
10 views

Counting the number of combinations from 5 equals sets

I have 5 sets with three elements each: A can be either 2,1 or 0 B can be either 2,1 or 0 . . E can be either 2,1 or 0 You can only pick one element from one of ...
1
vote
0answers
7 views

Deviation due to conditioning

Let $A$ and $B$ be random variables. Can we upper-bound the following expression? $$ \mathbb{E}\Big[\Big(\mathbb{E}[A|B] - \mathbb{E}[A]\Big)^2\Big] $$ The above looks classical research. However, I ...
1
vote
0answers
15 views

Training accuracy suddenly drops to zero after a while in Tensorflow?

I was training a ConvNet on CIFAR10 based on the code presented in "Deep MNIST for Experts" tutorial by TensorFlow. After around a few thousand epochs with training accuracy gradually increasing as ...
0
votes
0answers
6 views

Discovery and identification of parametric models

I would like to have some pointers regarding a framework combining model and parameter identification. I would like to "discover" a model from the data I have (chemistry application) and at the same ...
1
vote
0answers
3 views

Should I run variable selection within MICE for Multiple Imputation?

I have a dataset with around 100 variables, and I plan on getting multiply imputed datasets using the mice package in R. The literature I have read seems to advocate regressing each variable against ...
-1
votes
0answers
17 views

There are many statistical erros in this results and discussion. Can someone help me locate them? [on hold]

Results: Data Coding and Exploratory Data Analysis All data were entered into a single SPSS data file for analysis. Seven variables were coded: 1. Sex 2. Total amount of television viewing hours ...
0
votes
0answers
18 views

robust mixed models 3 time points (y-side) and high dropout

I am running the model in R: model = rlmer(Tau ~ tract_FA_avg + (1|Subject), data=long2) (robust as I have outliers) Since I am relatively new to mixed ...
0
votes
0answers
8 views

Tolerance interval vs Prediction interval, which one is wider?

Let say we have a 95% prediction interval, versus 95%/99% tolerance interval, which one is wider? A. PI always wider than TI B. TI always wider than PI C. Depends on sample you get Thanks!
0
votes
0answers
6 views

Applying exactly same WEKA filter on train and test data (What to use in setinputFormat traindata or test data) )

I am using WEKA for classification. I need to perform pre-processing before it. I want to do three thing , tf-idf conversion, normalization and discretization. But I want exactly same pre-processing ...
1
vote
0answers
28 views

Can an independent t-test be used on paired data when the pairing is unknown?

Suppose the effectiveness of a training course is examined, and performance of each individual in a group is taken both before and after, and the differences are compared in a paired $t$-test. Would ...
0
votes
0answers
7 views

Top Percentile Definition

Suppose I am in the top 10 percentile of a particular test. (I know what it means to be in the tenth percentile, but not the top tenth percentile) Does this mean I am in the top 10% of all ...
1
vote
0answers
13 views

How many boxes have balls?

I have N boxes, each of which may or may not contain a ball. I want to know how many boxes have balls by sampling the boxes in the following way. For each box in the N boxes: I open this box with ...
0
votes
0answers
9 views

How to calculate Mean and Standard deviation in Lognormal distribution where only P90 and P10 are known

I have P90 and P10 of a lognormal distribution. P90=142.56 and P10=3415; I need to estimate the mean and standard deviation of this lognormal distribution. I have done the following steps and then ...
-1
votes
0answers
15 views

Bivariate Poisson regression and Pythagorean Expectation for soccer(football) [on hold]

based on public data and using excel 2010 or after, I want to forecast/predict the football match winner.
0
votes
0answers
18 views

Optim error training ARIMA model in R [duplicate]

I have the code below which trains ARIMA models for a range of order combinations. I'm getting the error below in the step training the ARIMA models. The code worked just fine with the ...
0
votes
0answers
7 views

JAGS missing data problem

For missing data problem in RJAGS, in this site, there are a lot of posts talking about setting priors to generate data for the missing place. However, instead of filling up the places, I tend to ...
-1
votes
0answers
9 views

rmr2 rhadoop - outofmemory exception [on hold]

We trying rmr2 package from RHADOOP - get outofmemory exception. Below is sample code - aggregate sales quarterly. What is signature for function map, reduce & keyval? Some documents says: ...
2
votes
0answers
25 views

Why is the degrees of freedom for a matched pairs $t$-test the number of pairs minus 1?

I am used to knowing "degrees of freedom" as $n - r$, where you have the linear model $$\mathbf{y} = \mathbf{X}\boldsymbol{\beta} + \boldsymbol{\epsilon}$$ with $\mathbf{y} \in \mathbb{R}^n$, ...
2
votes
1answer
42 views

Resampling pre - post data by shuffling pre/post

Is the following approach a valid way to use resampling to determine if observed changes in pre/post test are significant: The null hypothesis is that there is no difference between the pre and post ...
4
votes
0answers
19 views

Recurrent Neural Network (RNN) topology: why always fully-connected?

I've started reading about Recurrent Neural Networks (RNNs) and Long Short Term Memory (LSTM) ...(...oh, not enough rep points here to list references...) One thing I don't get: It always seems that ...
0
votes
0answers
17 views

I have 5 models and each one consists of two 95 % confidence intervals, one for each axes, how can I plot these in R (ggplot2)? [on hold]

I have 5 models and each one consists of two 95 % confidence intervals, one for each axe, how can I find and plot these in R (ggplot2)? I tried to put the data in an ascending order and find them ...
1
vote
0answers
12 views

choosing prior parameters for variational mixture of Gaussians

I am implementing a vanilla variational mixture of multivariate Gaussians, as per Chapter 10 of Pattern Recognition and Machine Learning (Bishop, 2007). The Bayesian approach requires to specify ...
0
votes
0answers
10 views

xboost performance with predicted values as input

I have predicted probability of loss using different features. Now when I used this with non-important feature to predict probability of loss. It is very close first one. logloss was close to 0.11. ...
-1
votes
0answers
14 views

autoclassification in R [on hold]

Can you tell me, is there in R the possibility to choose the method that gives the most accurate classification well, for example ...
3
votes
1answer
31 views

Help with intuition about Chi-Squared distribution and its relation to Normal Distribution

At $\alpha = 0.05$, the significance cutoff of a Chi-squared distribution with 1 df is approximately 3.84, and that if we take the square root of 3.84, it's approximately 1.96, which is the 97.5 ...
0
votes
1answer
21 views

One side tests for example monte carlo

I looked at R for Spatial Statistics (spatstat) and I actually found out that it doesn't always have to be that below -1.96 is necessarily clustered and that above 1.96 must be uniformed. There is ...
1
vote
0answers
5 views

Method for Estimating Autocorrelation in Severely Gappy Data

My question is somewhat long and boils down to "Does the following work?" I'm working on a project that involves timing analysis of astrophysical data sets that have large chunks of data missing due ...
0
votes
2answers
15 views

Point Pattern Analysis- what is it good for

OK so I now know my point data is Random. Brilliant. But now what? How does it add anything to the understanding of the data? Sorry if that's a basic question but I honestly don't know what Point ...
0
votes
0answers
4 views

Planned Orthogonal Comparison Type I Error Rate and the Number of Independent Contrasts

I am confused about Type I error rate and the number of planned independent contrasts in Planned Orthogonal Comparisons in one-way ANOVA. If we have a factor with 5 groups and we would like to test ...
3
votes
1answer
11 views

Rao's Score/Lagrange Multiplier Test most powerful when $\theta$ close to $\theta_0$?

I was reading over the documentation for the Rao's Score/Lagrange Multiplier Test on wikipedia and ran across this: Rao's score test is a statistical test of a simple null hypothesis that a ...
0
votes
0answers
23 views

Can't Get a Distribution That's Reasonably Close to Data [on hold]

I've got about 30K data points. I've sampled about 1000 of them here (for reproduction's sake; I figured 1000 data points pasted here was probably a bad idea). ...

15 30 50 per page