Hi lir24,
What type of errors are you getting?
Martijn
Search found 76 matches
- Mon Jun 03, 2013 1:39 am
- Forum: SMILE
- Topic: smile embeded to iphone app
- Replies: 4
- Views: 5109
- Fri May 24, 2013 2:30 pm
- Forum: SMILE
- Topic: jsmile for IOS system?
- Replies: 1
- Views: 2674
Re: jsmile for IOS system?
Hi Lihua,
On our download page you can find a version for IOS under "Unix like"
Best,
Martijn
On our download page you can find a version for IOS under "Unix like"
Best,
Martijn
- Sat May 11, 2013 6:06 pm
- Forum: GeNIe
- Topic: Bayesian search and negative best score
- Replies: 5
- Views: 5938
Re: Bayesian search and negative best score
In this case we calculate classification accuracy for each network structure that we try and pick the one with the highest accuracy. This is not really a typical way of doing it, but useful when learning networks for classification. Here the score represents the number of correct classifications don...
- Sat May 11, 2013 3:09 pm
- Forum: GeNIe
- Topic: Bayesian search and negative best score
- Replies: 5
- Views: 5938
Re: Bayesian search and negative best score
No, that's a too simplistic view. Consider building a classifier that has a class variable and a few feature variables. Learning parameters for this model would give you a likelihood score. I can guarantee you that if you drop the feature variables and you learn parameters for just the class variabl...
- Thu May 09, 2013 5:12 pm
- Forum: GeNIe
- Topic: Bayesian search and negative best score
- Replies: 5
- Views: 5938
Re: Bayesian search and negative best score
Hi Ind,
The score is a loglikelihood.
Since the likelihood is a number between 0 and 1 the log likelihood will be negative.
For the loglikelihood, higher (in this case, closer to 0) is better.
Best,
Martijn
The score is a loglikelihood.
Since the likelihood is a number between 0 and 1 the log likelihood will be negative.
For the loglikelihood, higher (in this case, closer to 0) is better.
Best,
Martijn
- Thu May 09, 2013 5:10 pm
- Forum: SMILE
- Topic: Python for SMILE
- Replies: 7
- Views: 6849
Re: Python for SMILE
Hi,
We do not have a Python wrapper available.
Not sure if anyone ever created one on their own.
Best,
Martijn
We do not have a Python wrapper available.
Not sure if anyone ever created one on their own.
Best,
Martijn
- Thu May 09, 2013 5:08 pm
- Forum: SMILE
- Topic: SMILE API suggestion: load XDSL network from string
- Replies: 5
- Views: 5575
Re: SMILE API suggestion: load XDSL network from string
This has been included in the library.
Best,
Martijn
Best,
Martijn
- Fri May 03, 2013 2:39 pm
- Forum: SMILE
- Topic: setVirtualEvidence for C#
- Replies: 2
- Views: 3624
Re: setVirtualEvidence for C#
Hi, Did you read my email? Currently VirtualEvidence is not supported in the wrapper. We would have to add it, but this may take a little bit. You have the option of extending the wrapper yourself if you are really in a hurry. See http://genie.sis.pitt.edu/wiki/Appendices:_Extending_SMILE.NET_Librar...
- Mon Apr 08, 2013 10:08 pm
- Forum: GeNIe
- Topic: DBN, inference algorithms and unrolling
- Replies: 33
- Views: 25286
Re: DBN, inference algorithms and unrolling
Depends on your model structure, but assuming you set evidence for "evidence nodes" and are looking at the "hidden nodes", I'd say yes if you are looking a the hidden node x in time slice 5.
- Tue Apr 02, 2013 1:19 pm
- Forum: GeNIe
- Topic: Can't change the colour of arcs
- Replies: 4
- Views: 5391
Re: Can't change the colour of arcs
Hi Andrew, The grey color isn't really a new default color, it actually has a specific meaning. It means that the child node is in fact conditionally independent of the parent. If you mouse over the arc (the arrow head works best for this), you'll see it's labeled as unnecessary. If you make any cha...
- Tue Apr 02, 2013 2:49 am
- Forum: GeNIe
- Topic: DBN, inference algorithms and unrolling
- Replies: 33
- Views: 25286
Re: DBN, inference algorithms and unrolling
P(Zt|Zt-1....Zt-k)= *i=1-n P(Zit|Pa(Zit)) Murphy's algorithm works with DBNs that adhere to the first order Markov assumption, in GeNIe we allow for higher order networks, meaning that we allow temporal arcs to go further than just the next timestep. That's what this equation means. and the initial...
- Tue Apr 02, 2013 2:41 am
- Forum: GeNIe
- Topic: DBN, inference algorithms and unrolling
- Replies: 33
- Views: 25286
Re: DBN, inference algorithms and unrolling
No inference is ever performed on a rolled up DBN.saifjamal wrote:can i have any documentation on how the polytree working with the rolled up network?
We always unroll first
Martijn
- Mon Mar 25, 2013 8:37 pm
- Forum: GeNIe
- Topic: DBN, inference algorithms and unrolling
- Replies: 33
- Views: 25286
Re: DBN, inference algorithms and unrolling
The inference results will be the same. The difference is in the amount of necessary space needed for doing the inference. We unroll the network, which means we need space for the whole network. Murphy's algorithm just needs space for both time slices (the junction trees for the two slices, to be ex...
- Mon Mar 25, 2013 8:10 pm
- Forum: GeNIe
- Topic: DBN, inference algorithms and unrolling
- Replies: 33
- Views: 25286
Re: DBN, inference algorithms and unrolling
...where in the backward pass of polytree algorithm at time t will update the belief of the hypothesis node in the previous time slice. which is the cause they call it special case of HMM. You have to understand that you have to drop the references to HMM and it's algorithm. It's irrelevant. We unr...
- Mon Mar 25, 2013 1:36 pm
- Forum: GeNIe
- Topic: DBN, inference algorithms and unrolling
- Replies: 33
- Views: 25286
Re: DBN, inference algorithms and unrolling
is the first order temporal arc means that the relationship between two time slices modeled by first order hidden markov model. thats what made me confused A Hidden Markov Model is a special case of a Dynamic Bayesian network. It only has temporal arcs of order 1, meaning that nodes with temporal a...