HMM and CRF Package

[subhabangalore]

Dear Group,

I was looking for the following solutions.

(i) a Python Hidden Markov Model(HMM) library.
(ii)a Python Conditional Random Field(CRF) library.
(iii) I am using Python 3.2.1 on Windows 7(64 bit) and also like to get a NLTK version.
(iv) I may use unicode character as input.

If any one may kindly help me out.

Best Regards,
Subhabrata Banerjee.
Gurgaon.
India.

 

[subhabangalore]

Dear Group,

I was looking for the following solutions.

(i) a Python Hidden Markov Model(HMM) library.
(ii)a Python Conditional Random Field(CRF) library.
(iii) I am using Python 3.2.1 on Windows 7(64 bit) and also like to get aNLTK version.
(iv) I may use unicode character as input.

If any one may kindly help me out.

Best Regards,
Subhabrata Banerjee.
Gurgaon.
India.

Dear Group,
I worked out a solution. As most of the libraries give so many bindings andconditions best way is to make it. Not very tough, I made earlier, but as some files were lost so was thinking instead of a remake if ready versions work. Or may look change from Python 3 to previous version so when 3 will be supported solution. But can I install both Python 3 and Python2.x or call 2.x from 3.x?

Regards,
Subhabrata.

 

[Terry Reedy]

As most of the libraries give so many bindings and conditions best way

is to make it. Not very tough, I made earlier, but as some files were
lost so was thinking instead of a remake if ready versions work. Or may
look change from Python 3 to previous version so when 3 will be
supported solution. But can I install both Python 3 and Python2.x or
call 2.x from 3.x?

You can have multiple versions of Python installed. Just be careful on
*nix to use altinstall so as to not disturb the system install used by
the system.

Pythonx.y can only use code that works with x.y ;=). However, any Python
process can use any other program via the subprocess module. You just
have to work out the data transfer.

 

[subhabangalore]

On 7/25/2012 11:58 AM, (e-mail address removed) wrote:
> As most of the libraries give so many bindings and conditions best way
is to make it. Not very tough, I made earlier, but as some files were
lost so was thinking instead of a remake if ready versions work. Or may
look change from Python 3 to previous version so when 3 will be
supported solution. But can I install both Python 3 and Python2.x or
call 2.x from 3.x?

You can have multiple versions of Python installed. Just be careful on
*nix to use altinstall so as to not disturb the system install used by
the system.

Pythonx.y can only use code that works with x.y ;=). However, any Python
process can use any other program via the subprocess module. You just
have to work out the data transfer.

Dear Terry,
Thanks for your reply. I had downloaded it too, but going tools I felt it is better to repair or rebuild them, as most of them have many bindings. And making is not a big effort when numpy,scipy,matplotlib are there.

Regards,
Subhabrata Banerjee.

 

[subhabangalore]

On 7/25/2012 11:58 AM, (e-mail address removed) wrote:
> As most of the libraries give so many bindings and conditions best way
is to make it. Not very tough, I made earlier, but as some files were
lost so was thinking instead of a remake if ready versions work. Or may
look change from Python 3 to previous version so when 3 will be
supported solution. But can I install both Python 3 and Python2.x or
call 2.x from 3.x?

You can have multiple versions of Python installed. Just be careful on
*nix to use altinstall so as to not disturb the system install used by
the system.

Pythonx.y can only use code that works with x.y ;=). However, any Python
process can use any other program via the subprocess module. You just
have to work out the data transfer.

Dear Terry,
Thanks for your reply. I had downloaded it too, but going tools I felt it is better to repair or rebuild them, as most of them have many bindings. And making is not a big effort when numpy,scipy,matplotlib are there.

Regards,
Subhabrata Banerjee.