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

learn more… | top users | synonyms (1)

3
votes
0answers
19 views

Is it possible for a reinforcement learning agent to create or generate additional features

I have little background knowledge of Machine Learning, so please forgive me if my question seems silly. Based on what I've read, the best model-free reinforcement learning algorithm to this date is ...
2
votes
2answers
45 views

forecasting sharp seasonal peak in time series

I have time series data on a daily level over the past 4 years. What is clear from examining past data is that there are two very clear peaks in the time series around the same time of year (they ...
3
votes
2answers
67 views

Generating recommendations using matrix multiplications

The Mahout In Action (Chapter 6) book contains a recommendation method based on matrix multiplication that uses co-occurrence data (C) in combination with user preferences (U) to generate user ...
0
votes
1answer
25 views

Dimension Reduction

I have a n*m matrix, the rank of matrix (r) is near to min(m,n) I want to minimize the rank by removing some of the rows or columns to get r << min(m,n) The goal is to achieve least rank for ...
0
votes
1answer
35 views

How many data do you need for a convolutional neural network?

If I have a convolutional neural network (CNN), which has about 1,000,000 parameters, how many training data is needed (assume I am doing stochastic gradient descent)? Is there any rule of thumb? ...
1
vote
0answers
16 views

What is “fitted function” in the context of boosted regression tree?

I'm following the tutorial of package dismo's boosted regression tree, which produces two graphs, about fitted function and ...
3
votes
0answers
25 views

Bounding the expectation of the difference between empirical vs generalization error

Let the (defect) difference between empirical and generalization error be: $$D[f_S] = I_S[f_S] - I[f_S]$$ where the empirical risk is: $$I_S[f_S] = \frac{1}{n}\sum^n_{i=1} V(f_S,z_i)$$ where ...
3
votes
0answers
35 views

Explanation for large difference in SVM and Naive bayes results

I have a dataset with 389 data evenly distributed into 6 classes. Each data has 1024 features. So my dimension is much larger than my observation data. I have tried to see some common classifiers on ...
1
vote
1answer
28 views

Questions about Q-Learning using Neural Networks

I have implemented Q-Learning as described in, http://web.cs.swarthmore.edu/~meeden/cs81/s12/papers/MarkStevePaper.pdf In order to approx. Q(S,A) I use a neural network structure like the following, ...
0
votes
1answer
61 views

What are the benefits and disadvantages to Lasso, Ridge, Elastic Net, and Non Negative Garrotte Regularization techniques?

I am implementing these four regularization techniques for linear regression of stock data in MATLAB but i noticed elastic net is just the sum of Ridge and Lasso, and i dont full understand how ...
3
votes
1answer
46 views

How do you find mathematical expressions for the posterior marginals i.e. $P(x_n|y_0, … , y_n)$ in an HMM?

My goal is to find closed form equations for posterior marginals $P(x_n|y_0, ... , y_n)$ in a general HMM. I was told that we can calculate it exactly via BP (belief propagation, thought not sure how ...
4
votes
1answer
99 views

Justifying and choosing a proper scoring rule

Most resources on proper scoring rules mention a number of different scoring rules like log-loss, Brier score or spherical scoring. However, they often don't give much guidance on the differences ...
1
vote
1answer
16 views

How do I view “error correct learning” in ANN as an optimal control problem?

There is a lot of material out there for the gradient descent method used in ANN but no body makes it clear how this is an optimization problem or brush it off as extraneously info. Can someone make ...
1
vote
2answers
68 views

SVM heavily over fits the data (classifying Highly Unbalanced data )

I have a huge training set from which I am supposed to regress and classify, i.e I classify whether an event will occur or not and another task is to regress the intensity of the event in future. The ...
1
vote
1answer
35 views

Cross validation in semi-supervised learning

With semi-supervised learning a labeled set $X_L$ and unlabeled set $X_U$ are given. If the learning algorithm has several free-parameters we are forced to perform cross-validation to try to guess ...
0
votes
1answer
24 views

Association rules or classifier for product modeling for queries

I have a set of products P {1...n} which are rated on a goodness scale G ={1...100} (G10 is more good than G5). Each product has a set of features F {1....m}, now I want to learn a model for ...
3
votes
1answer
43 views

SVM: Number of support vectors

Imagine I am using an SVM to train a classifier for a given dataset, in one-vs-all configuration. For each class, I am performing cross validation for parameter selection (grid search to choose ...
0
votes
1answer
22 views

Novelty and Outlier Detection for Multi-label Data

I met a problem of using novelty and outlier detection for my multi-label data. For example, I have got some training data that is not polluted by outliers. However, the training data are with ...
1
vote
1answer
75 views

Is a neural network the only way to learn input/output relation?

first question here but hopefully you can help. Firstly I'm a web programmer by trade but I've touched a bit of neural networks in my university days. I've got a project to do with predicting the ...
2
votes
1answer
58 views

A valid distance metric for high dimensional data

I asked a question about forming a valid distance metric yesterday (Link1) and got some very good answer; however, I have got some more questions about forming a proper distance metric for high ...
0
votes
0answers
13 views

pybrain LSTM layer buffer variables

In pybrain LSTM layer there are these buffer that are used to store values. ...
0
votes
0answers
33 views

Improving dynamic time warping word recognition system

I recently got interested in speech recognition and have implemented a simple dynamic time warp system for word recognition for my own learning purpose. However after testing a bit I believe that I ...
4
votes
1answer
144 views

Dealing with hierarchical (panel, multi-level) data and fixed effects in LASSO?

The question pretty much explains itself. When running a Lasso regression on a lot of indexed (say by time and location) explanatory variables, is it best practice to transform all data using a ...
0
votes
1answer
15 views

Nonlinear Dynamic Online Classification: Looking for an Algorithm

I have two predictors a,b that I want to use combine to classify data. a is stable, it will always produce the same prediction for the same input. b will change and probably improve in time (because ...
3
votes
1answer
48 views

Time Series Forecast: Convert differenced forecast back to before difference level

I am using R and I need an easier way to produce forecasts of data at the original level based on forecasts using differenced data. The situation, in more detail, is this: I am using several ...
1
vote
1answer
27 views

Method for solving problem with variable number of predictors (repost from Data Science)

REPOST from Data Science: I've been toying with this idea for a while. I think there is probably some method in the text mining literature, but I haven't come across anything just right... What ...
0
votes
0answers
7 views

DecisionTreeClassifier scikit-learn : knowing the leaf to which an example belongs to

I am currently reading this paper http://quinonero.net/Publications/predicting-clicks-facebook.pdf , where they are using trees to generate feature that are afterwards fed to a linear classifier. My ...
1
vote
3answers
58 views

How to get a valid distance metric?

I have got a problem to devise a distance metric to get the similarity measurement of vectors. Someone suggested me to use dot product, which seems to me the same as the Cosine similarity metric; ...
0
votes
0answers
11 views

Full batch backpropagation implementation

I am trying to wrap my head around using batch backprop in a neural network. I have a very code-oriented mind, and I'm trying to figure out whether it's possible to parallelize the full batch ...
1
vote
1answer
49 views

Sparse coding vs. sparse PCA, are they the same thing?

Are they the same thing? If not, could someone possibly explain the difference or point to the seminal papers describing the approaches? I am looking not for a detailed technical exposition, but a ...
0
votes
0answers
14 views

Binary classifier issues

I am trying to predict if sales are going up or done given a specific set of features. The only thing I care about is precision here. In this context I tried a few classifiers ( SVM, Random Forest ) ...
0
votes
0answers
13 views

What is the “standard reference” for cascade forward neural network?

What is the "standard reference" that firstly describes or surveys in details the cascade forward neural network? This kind of net is available in matlab toolbox for long cascadeforwardnet (as early ...
1
vote
0answers
15 views

serial correlation, homocedasticity tests for non linear and non parametric regression

For linear and parametric regression there are multiple tests where variables and residuals are used by means of performing a linear regression function to test serial correlation of regression errors ...
0
votes
1answer
26 views

Biasing SkLearn Algorithms to Positive Outcomes

I am trying to run multinomial naive bayes on a series of examples in python using sci kit learn. I am consitently getting all examples classified as negative. (The ratio of positives to negatives in ...
0
votes
0answers
16 views

Calibrated Probabilities for Outlier Detection Algorithm

I am currently developping a 2 class classification algorithm. However, as the dataset is at the moment really small (<50 observations) and imbalanced (~1/10 ratio), I decided to rather first ...
3
votes
1answer
51 views

Which algorithm should I use?

I was reading many machine learning questions like this one but I am not sure how to apply them to my scenario. I come from a biology/medicine background, and my math knowledge is limited (last thing ...
1
vote
1answer
37 views

What are the advantages of ReLU over sigmoid function in deep neural network?

The state of the art of non-linearity is to use ReLU instead of sigmoid function in deep neural network, what are the advantages? I know that training a network when ReLU is used would be faster, and ...
3
votes
1answer
43 views

Optimal construction of day feature in neural networks

Working on regression problem I started to think about representation of "day of a week" feature. I wonder which approach would perform better: one feature; value 1/7 for Monday; 2/7 for Tuesday... ...
0
votes
1answer
35 views

Distribution Assumptions in Ridge & Lasso Regression Models?

What are the assumptions for the distribution of the features for regression models like Lasso regression or Ridge regression? Why is it better to have features with Gaussian distributions?
4
votes
1answer
33 views

Is there such thing as regression involving a pairwise response variable? (X,Y)~Z0+Z1*B1

I'm trying to model a pairwise outcome of basketball game scores. Ie. (94,87),(102,98),(76,54),... My input variable is a single performance metric for each team. Ie. (12,9),(14,17),... Is there a ...
0
votes
0answers
35 views

credit scoring - fraud scoring

I have been asked to build a credit scoring model and we are relying on several Machine Learning API, in order to build our feature vectors. One of these API is MinFraud. However, as they provide us ...
1
vote
1answer
20 views

Convolutional neural network - Using absolute of tanh on convolution output

I've watched an online lecture regarding CNN (https://www.youtube.com/watch?v=wORlSgx0hZY) that confused me a bit. At roughly 8:35 in the lecture it was stated that it is important to use the absolute ...
0
votes
0answers
12 views

R: how to cluster data by EM algorithm and get Mahalanobis distance?

I have some data, the format as follows: ...
0
votes
1answer
34 views

If random forests gives me a bad cross-validation score, should I trust it for feature selection?

I get an R^2 value of about 0.22 when I 10-fold cross-validate with my entire dataset. My main use for random forests is to analyze feature interactions. But should I trust the feature importances ...
0
votes
0answers
16 views

Edge prediction from Live Journal data in SNAP

I have downloaded data for the Live Journal graph from SNAP . The dataset looks like this From_Node_Id ->To_Node_Id ...
0
votes
0answers
11 views

Complement naive bayes

I would like some help in understanding how does Complement Naive Bayes work. I have googled the paper Complement Naive Bayes I understand that naive bayes works by computing the probability of a ...
0
votes
0answers
5 views

Is there any R programs can get the Row and Column space of a matrix? [migrated]

In R, there is a method Null which can get the null space of a matrix. But, is there any R code can get the row and column space of a matrix ?
0
votes
0answers
24 views

Machine Learning: Potential Reasons of Precision Change after New Features are Added

My baseline model uses 10 features $[f_1, f_2, \dotsb, f_{10}]$. Now I have two new features $f_{11}$ and $f_{12}$. New models that use either $[f_1, f_2, \dotsb, f_{10}, f_{11}]$ or $[f_1, f_2, ...
1
vote
0answers
9 views

Can summation of gradient descent runs on subsets approximate gradient descent run on set?

I have a training set M, which I then split into two subsets, say, M1 & M2. If I run gradient descent on each subset to come up with a linear regression which approximates the data in M1 & ...