Search found 44 matches
- Sat Mar 15, 2008 8:29 pm
- Forum: GeNIe
- Topic: Very long training Time with smaller network
- Replies: 39
- Views: 39986
Hello Marc, what do you exactly mean by using a prior? Could you please give me alittle example? I have an other idea what the reason for the long develope time could be. I have 3 times Genie open and I am training a network on each Genie instance as I have a quad core and need much networks to buil...
- Fri Mar 14, 2008 11:06 pm
- Forum: GeNIe
- Topic: Very long training Time with smaller network
- Replies: 39
- Views: 39986
I had the problem that sometimes states in the node with the many states showed 0 due to lack of available information I think. So I have put each three logical states to one. I thought I would get better results this way. Calculation speed does not make such a difference. I just thought when it is ...
- Fri Mar 14, 2008 10:19 pm
- Forum: GeNIe
- Topic: Very long training Time with smaller network
- Replies: 39
- Views: 39986
I am not that good in understanding the EM-Algorithm, but what you say sounds reasonable. As in the slower network the parameters are much more fluctuating, do you think the faster trained network will result in better forecasts as the values are more clear (not fluctuating so much) and so more corr...
- Fri Mar 14, 2008 2:41 pm
- Forum: GeNIe
- Topic: Very long training Time with smaller network
- Replies: 39
- Views: 39986
The network it a very big one and also the data is confidential. I hope you could solve the problem without seeing the data. But I could create a similar network with the same nodes (renamed), same connections and same states. I would have no training djata for this one. But as I could just rename e...
- Thu Mar 13, 2008 7:07 pm
- Forum: GeNIe
- Topic: Very long training Time with smaller network
- Replies: 39
- Views: 39986
Very long training Time with smaller network
Hello, I have build a network with 6,500,000 parameters, 13 nodes, 217 states. I let it train with 2,500,000 rows of training data. It needed 110 minutes. Thats okay. To make the net a bit smaller I used the complete same network, but joined each three 3 states in node one to one state. Every other ...
- Fri Jan 11, 2008 2:36 pm
- Forum: GeNIe
- Topic: Genie has problems with big nets
- Replies: 9
- Views: 12408
- Sat Jan 05, 2008 2:26 pm
- Forum: GeNIe
- Topic: Genie has problems with big nets
- Replies: 9
- Views: 12408
- Tue Dec 25, 2007 1:17 am
- Forum: SMILE
- Topic: DataSet.AddRecord(rec) not available in DotNet Build?
- Replies: 1
- Views: 5065
DataSet.AddRecord(rec) not available in DotNet Build?
Hello, I am just messing with the DotNet Smile wrapper build, but I found no possibility to add records to a dataset beside using ReadFile. Is there really no way? I think writing a Textfile and reading it again is not so good. Maybe there is a way and I haven't found it. In C++ there is a AddRecord...
- Sun Dec 16, 2007 5:53 pm
- Forum: GeNIe
- Topic: Genie has problems with big nets
- Replies: 9
- Views: 12408
Hello, sorry for my late answer, I was out of town lately. I have build an example for you. To reproduce the error, do the following: 1) Open the network 2) Connect Node6 to Node 1. 3) Press F5 to calculate. This should work fine every time. 4) Press Save As and save the network. Most of the time Ge...
- Wed Dec 12, 2007 9:17 pm
- Forum: SMILE
- Topic: Maximum net size / Multi threading
- Replies: 2
- Views: 6087
Maximum net size / Multi threading
Hello, what is the maximum network size I can use? Is it only limited by the 2GB border or also by a maximum number of nodes/weights/arcs? Is there also a multi thread calculation algorithm? I am currently using the Likelihood algorithm with 10k samples each on 4 cores. After that I calculate the av...
- Wed Dec 12, 2007 9:14 pm
- Forum: GeNIe
- Topic: Genie has problems with big nets
- Replies: 9
- Views: 12408
Genie has problems with big nets
Hello,
if I build a network with more than average 25 million values I can calculate, edit all the values and so on. But as soon as I want to save this net, GeNIe is crashing.
So it seems that smile is ok with that big networks, but GeNIe not?
if I build a network with more than average 25 million values I can calculate, edit all the values and so on. But as soon as I want to save this net, GeNIe is crashing.
So it seems that smile is ok with that big networks, but GeNIe not?
- Thu Dec 06, 2007 2:50 pm
- Forum: SMILE
- Topic: DSL_ALG_BN_LAURITZEN and DSL_ALG_BN_LSAMPLING
- Replies: 1
- Views: 5208
DSL_ALG_BN_LAURITZEN and DSL_ALG_BN_LSAMPLING
Hello, I am just trying the API, but have problems to figure out the exact name of the algorithm in relation to the used constants. I looked in the C# and C++ help I think I figured out that DSL_ALG_BN_LAURITZEN is the clustering algorithm. Is DSL_ALG_BN_LSAMPLING the logic sampling or the likelihoo...
- Thu Dec 06, 2007 7:46 am
- Forum: GeNIe
- Topic: Continuous variables
- Replies: 23
- Views: 28390
ahh, okay - I understand. Thank you very much for your detailed examples. Thats a good approach. Then that's the way to go for me. But if I need soft evidence in a parent node I can just change the weights? For example in your net1.jpg - Node1, if I want 60% state0 and 40% state1 I can just set it, ...
- Wed Dec 05, 2007 9:22 pm
- Forum: GeNIe
- Topic: Continuous variables
- Replies: 23
- Views: 28390
I am trying to simulate a soft evidence. For example if I have a node with 60% state0 and 40% state1 I build 10 examples, 6 with state0 to 100% and 4 with state1 to 100% After training this net I should have the 60%/40% for that node. Now let asume I have two nodes that are connected (Node 1 is pare...
- Wed Dec 05, 2007 5:57 pm
- Forum: GeNIe
- Topic: Continuous variables
- Replies: 23
- Views: 28390