Methods and principles of building "computer systems that automatically improve with experience."

learn more… | top users | synonyms (1)

1
vote
1answer
33 views

Ensembles - few questions on approach with multiple models

I'm looking for some general-high level understanding of how I should apply ensemble techniques. I understand that some models can be already thought of as ensembles - such as random forests or some ...
0
votes
1answer
21 views

Normalising/Standardising Data for Machine Learning

If we have a system in which we normalise and standardise the data, I'm interested into how to most effectively do this. For a test system we can apply normalisation and standardisation techniques to ...
3
votes
1answer
51 views

ML model selection for prediction of latitude and longitude

I am doing a project in which my aim is to predict the likely locations of a set of latitude/longitude points based on a couple of variables. Since I've never done any ML on locational data, which ...
3
votes
0answers
36 views

Can we use gradient desent method in maximum entropy model?

I see a lot of implementations use GIS or IIS to train the maximum entropy model. Can we use gradient desent method? If we can use it, why most tutorial directly tell GIS or IIS methos, but do not ...
0
votes
0answers
4 views

Baseline for dummy variables in XGB. Is it important?

I have a lot of categorical variables that I am using in an XGB model. I use the package dummies to make dummy variables and it creates dummies for all levels. There is no baseline. I know that in ...
2
votes
1answer
32 views

Determine the feature weights with a regression?

I have a set of houses with their features (location, size, number of rooms, etc … and the y is the price). In the future I will have a new house without the price. My goal is to find the 20 closest ...
0
votes
1answer
38 views

How to make sense of and use error derivatives of backpropagation algorithm

I'm following Geoffrey Hinton's Neural Networks Lectures (this one in particular: https://www.youtube.com/watch?v=LOc_y67AzCA), and I'm stumped by the second equation here: As I understand it, $E = ...
0
votes
0answers
25 views

Why does executing a PCA on my unbalanced dataset change the prediction distribution of my test so much?

I'm using prtools to try and solve a machine learning problem and trying to use a PCA to help me with this. Now I have a dataset with 13 features (columns), 10000 ...
0
votes
1answer
19 views

How to compute logarithmic loss with the probability output of binary classification

The following code are used to produce the probability output of binary classification with Random Forest. ...
0
votes
1answer
48 views

How to find the best clustering method?

I have a data set where the samples are people and the feature are their age, sex, location, job, height, weight… Then I will have a new person with the same information and my goal is the find the ...
3
votes
1answer
53 views

Eigenvalue equation for kernel PCA

In Nonlinear component analysis as a kernel eigenvalue problem, Schölkopf et al start by describing PCA. Given a set of data instances $x_1, \dots, x_M$, with $x_k \in \mathbb{R}^N, k=1,\dots,M$, and ...
0
votes
0answers
12 views

What is the meaning of output in weka- SMOreg regression?

I am new to Weka. I was trying SMOreg for a regression on their example data file 'cpu.arff' . I got this output : ...
9
votes
3answers
209 views

Why are neural networks becoming deeper, but not wider?

In recent years, convolutional neural networks (or perhaps deep neural networks in general) have become deeper and deeper, with state-of-the-art networks going from 7 layers (AlexNet) to 1000 layers (...
1
vote
1answer
12 views

How NN processes a pattern\ dataset formed of millions of points as a single input?

I would like to know how NN processes a pattern\ dataset formed of millions of points as a single input. Does NN cluster the dataset and or extract the features? How does Sigmoid transfer function ...
0
votes
4answers
47 views

Can you use neural networks for non-binary input feature sets?

I want to feed a set of non-binary features/attributes of my problem into the input for a neural network. Currently, I am looking about 21 features I would like to use for my input and one binary ...
2
votes
1answer
83 views

Comparing SVM Models using Different Methods for Data Generation

I have a set of SVM models that I am trying to compare. Each of the models is trained on a variation of the original data: The original data The original data using resampling scheme A The original ...
4
votes
4answers
134 views

Should I get 100% classification accuracy on training data?

I've been getting inconsistent results with a binary classification problem I'm trying to solve using a linear classifier and a custom feature extraction pipeline, and decided to do a quick check of ...
1
vote
0answers
18 views

Getting prediction intervals from GBM models

I am working with GBM regression models(in H2O) and am using Quantile distribution for the distribution parameter. I am looking for a method to provide prediction intervals in addition to point value ...
0
votes
0answers
25 views

Maximizing the likelihood function for logistic regression

I've been working with logistic regression and watching tutorials how to maximize the likelihood function $$\prod_{i=1}^n P(x_i)^{x_i}(1-P(x_i))^{1-x_i}$$ According to the sources I've been watching, ...
1
vote
1answer
84 views
+50

Can I hack weighted loss function by creating multiple copies of data

Suppose we want to build a binary classifier with weighted loss, i.e., it penalize different types of errors (false positive and false negative) differently. At the same time, the software we are ...
1
vote
1answer
24 views

SVM labeling more classes does not result in higher accuracy?

I used SVM to predict the ranking score of muffin recipes. X is a numpy array of ingredient amounts of a certain recipe and y is the label according to the online ranking score. First I labelled my ...
1
vote
0answers
14 views

Chi square kernel function: which version?

I'm not a machine learning expert, so sorry if this is a stupid question. I was reading this paper about Kernenlized locality-sensitive hashing (LSH) and in section 6.4 the chi-square-kernel function ...
0
votes
1answer
28 views

Regression models similar to Random Forest

I am trying to create an ensemble model. I tried some classical models, in which RandForest's accuracy is around 85% and other ...
0
votes
0answers
19 views

Number of Sinusoids Feature Extraction for MUSIC Algorithm

I am a complete novice to machine learning, but I wanted to ask this question to get me started on a path to solving this problem. I have been working on a radar project where I am receiving a signal ...
0
votes
0answers
21 views

Machine Learning methods for classification with continuous variables

I'm looking for suggestions for some methods for classification (3 classes) using continuous input variables. I'm aware of using neural networks for this purpose but most of the other ML recipes I'm ...
3
votes
1answer
41 views

a simplified version of fully convolutional network

In the paper of fully convolutional networks semantic segmentation, the authors adopts up-sampling (de-convolutional network) to recover the feature maps with reduced dimensions (due to the multiple ...
-1
votes
3answers
37 views

How to group/cluster similar words

Let say I have a list of words, such as: apple apale aaple apples oranges ornnges orange orage melons meeons meeon melon melan I want to group them based on ...
0
votes
0answers
13 views

Different logistic regression models

I've been revising the concepts of logistic regression and suddenly realized that the probability function of the logistic model in the book ISL looks absolutely different from other sources. For ISL ...
1
vote
0answers
12 views

Right procedure for Leave-One-Out Cross-Validation in multi-class classification

I am intending to implement a existent method to recognize sign language. The model uses a data-set of N (375) classes but with only 3 examples per class. The authors used Leave-One-Out Cross-...
3
votes
0answers
68 views

Splitting data for train/test for time series

A week ago or so I was at a conference. Long story short, I ran into a friend who is quite good at machine learning so I asked them a question about why I might be getting what I think is poor fit on ...
3
votes
1answer
39 views

Why is training a deep convolutional neural network taking longer time than anticipated?

It's taking me over 4 days to train a deep learning network with just 10000 images of 224px x 224px x 3 channels size, with batch size 25. The machine has 32GB RAM, a Core i7 CPU, and a GTX 960 GPU. I'...
4
votes
0answers
62 views

What are the impacts of choosing different loss functions in classification to approximate 0-1 loss

We know that some objective functions are easier to optimize and some are hard. And there are many loss functions that we want to use but hard to use, for example 0-1 loss. So we find some proxy loss ...
0
votes
2answers
89 views

random forest how to use the results

I used the package for random forest. It is not clear to me how to use the results. In logistic regression you can have an equation as an output, in standard tree some rules. If you receive a new ...
1
vote
1answer
45 views

How is the root node of a decision tree determined?

Almost all the examples I have found stated how the decision tree's split is based on how much purity/information can be gained (ie: via entropy and information gain) for internal node. But is the ...
2
votes
0answers
21 views

What does the term data dependent kernel mean?

I was reading a paper that revisits (singled layered) Radial Basis Function networks. In they claim that they show that RBF methods can be recast as "certain data-dependent kernels". I was wondering ...
0
votes
0answers
21 views

Feature map of kernel

Let $K_1,K_2$ be valid kernels. Kernel $K_1$ has a feature map $Θ(x)∈R^{50}$ while Kernel $K_2$ corresponds to feature map $ψ(x)∈R^{10}$, satisfying: $∀i=1..10,∀x:ψ_i (x)=0.2 Θ_i (x)$ What is the ...
0
votes
0answers
29 views

Smart sampling techniques to ensure boundary/edge cases are included

In an effort to reduce the time it takes to test software against various data sets I would like to do create an application that looks at a data set and creates a "smart sample" from the it. The ...
0
votes
0answers
17 views

How to use weights with Elasticnet regression in python?

I am using Elasticnet from scikit-learn in python, I've also used Glmnet package in R for prototyping. I want to use weights in Elasticnet which apparently is not available as an option/argument in ...
1
vote
0answers
58 views

Using PCA to model highly correlated variables

Specifically, Andrew Ng states that PCA should be used to speed up algorithms or to visualize data. He also states that using PCA as a way to prevent overfitting is an incorrect application of PCA. ...
0
votes
0answers
24 views

Train/test split for a small dataset (classification tree/ random forest)

I'm trying to solve a classification problem using Classification Tree with a small amount of data available. Depending on the content of particular train/test subsets (80:20 proportion) the total ...
0
votes
0answers
11 views

NEAT Neural Network Digit Recognition Mean Square Error 0.9

I'm new to machine learning and am trying to create a neural network using the NEAT algorithm in python that can recognize 0-9 digits from the sklearn data set. There are ten output neuron in the nn, ...
1
vote
1answer
21 views

Same kernel for mixed/categorical data?

I know it's common practice, but is it right to apply the common kernels to categorical/mixed data? If not, are there alternatives? I'm expecting answers from both theoretical and practical points of ...
1
vote
0answers
29 views

Is there difference between “spectral decomposition” and “singular value decomposition”? [duplicate]

Am I right that "spectral decomposition" for symmetric matrix and "singular value decomposition" for non square matrix? Any clarification would be appreciated.
4
votes
2answers
67 views

Should random forests based on same data but different random seeds be compared?

I know that if you re-run a random forest with a different random seed you will fit a different model. I'm wondering whether it's acceptable to compare different random forest models (run under ...
0
votes
0answers
15 views

Quadratic programming in SVM is inefficient in higher dimension?

I am learning SVM from this lecture - https://www.youtube.com/watch?v=eHsErlPJWUU - though I have some questions. Let's say we have d features: x1, x2, x3..., xd. And we have p sample: x1(1), x2(1), ...
1
vote
0answers
25 views

improving classification using examples from 'other' classes

I'm trying to classify a a small dataset of around 150 patients each with one of four different diseases. For each patient I have around 70 different features. I also have an "extra" dataset of ...
0
votes
1answer
76 views

why cost function needs to be smooth, how this helps in learning?

From what I understand a smooth function 1-degree is a function whose first derivative is continuous? How this helps in estimating the parameters? What if the function is not smooth?
0
votes
0answers
16 views

Best practices to evaluate semi-supervised learning methods

Some background I have been working on multiclass classification method. I have an idea on how to extend this method such that it can be semi-supervised. What are the best practices in evaluating ...
1
vote
1answer
9 views

Why provide links as input to relational topic model?

I am studying relational topic models, which are designed to model the topics in a corpus similar to LDA as well as the links between them. However, it seems like an input to this algorithm is in ...