# Nick

less info
reputation
1717
bio website location CA age member for 3 years, 1 month seen Nov 4 at 22:51 profile views 89

I'm a computer scientist working with lots of data! :)

# 141 Actions

 May30 comment Proposal for transition matrix for Metropolis-Hastings phylogenetic inference You say that the substitution probabilities are not required to sum to 1, are they independent? is the sum required to be <= 1? May26 comment Problem of “clustering” into most similar groups Could you get to the same result by first clustering with k-means, then constructing your 'groups' by taking M points from each cluster? May25 comment Classification-results datasets what exactly is your question? How to get the classification results? May25 revised How to calculate threshold level for mutual information scores? grammar May24 answered How to calculate threshold level for mutual information scores? May23 comment Observation (evidence) in dynamic Bayesian networks observed nodes in a BN are nodes for which we have seen the actual values (e.g. from sensor data). The value has a likelihood given the node parameters. For instance, if the observed value from a sensor is assumed to have Gaussian noise about the true value and our current estimate of the true value is $x_t$, then the observed sensor value: $z_t$ has a likelihood of: $\mathcal{N}(z_t; x_t, \sigma)$. May23 comment Observation (evidence) in dynamic Bayesian networks This question is rather unclear. Are you asking how the observed nodes in a Bayesian network affect the probability of the unobserved nodes? How to model the values for your specific problem (i.e. what distributions to use)? Clarifying your question will greatly increase our ability to provide a relevant answer. May22 comment How to cluster LDA/LSI topics generated by gensim? You can treat the topic mixture vector for each document as its position in this latent topic space. Simply run your clustering using this as the input data. What is the point of the hard clustering though? The topic mixtures already give you a lot of information about how documents are similar/different. May22 revised What is the difference between “priors” and “likelihood”? minor notation change May22 comment What is the difference between “priors” and “likelihood”? @NeilG, yes, you're absolutely correct - I've edited my answer to reflect this. Its sometimes written $\mathcal{L}(\theta | D)$ because the likelihood can be viewed as a function of the parameters holding the data constant. May19 answered What is the difference between “priors” and “likelihood”? May17 comment Constructing a Bayesian network from the begining You can use maximum likelihood estimates (this results in simply counting # of transactions that include fraud = yes, normalized by the total number of samples to estimate p(fraud = yes/no)). May16 answered Constructing a Bayesian network from the begining May2 comment matlab princomp latent can you post the exact code you used to run PCA? Apr23 comment MCMC algorithm to estimate beta and variance @Matt: are you talking about understanding how to go about setting up and computing Bayesian inference for a linear regression model? If so, please edit your question to say that, if not, please explain what you mean. Apr10 awarded Enthusiast Apr7 comment Choosing attributes for clustering/classification Can you provide any context about the type of data this is and why you only have 2 instances for each class (I'm just curious)? Apr6 comment Dealing with lots of ties in kNN model How do you get ties if $k$ is odd? What do you mean by "only using variables with >2 levels"? Apr2 comment Is Bayesian nonlinear regression using conjugate priors possible? Depends on the sampling distribution (likelihood). A table showing the conjugate prior of various likelihood models is shown on wikipedia: en.wikipedia.org/wiki/Conjugate_prior Apr2 answered Simple introduction to MCMC with Dirichlet process prior?