function?[net?options]?=?glmtrain（net?options?x?t）
%GLMTRAIN?Specialised?training?of?generalized?linear?model
%
%	Description
%	NET?=?GLMTRAIN（NET?OPTIONS?X?T）?uses?the?iterative?reweighted
%	least?squares?（IRLS）?algorithm?to?set?the?weights?in?the?generalized
%	linear?model?structure?NET.??This?is?a?more?efficient?alternative?to
%	using?GLMERR?and?GLMGRAD?and?a?non-linear?optimisation?routine
%	through?NETOPT.?Note?that?for?linear?outputs?a?single?pass?through
%	the??algorithm?is?all?that?is?required?since?the?error?function?is
%	quadratic?in?the?weights.??The?error?function?value?at?the?final?set
%	of?weights?is?returned?in?OPTIONS（8）.?Each?row?of?X?corresponds?to
%	one?input?vector?and?each?row?of?T?corresponds?to?one?target?vector.
%
%	The?optional?parameters?have?the?following?interpretations.
%
%	OPTIONS（1）?is?set?to?1?to?display?error?values?during?training.?If
%	OPTIONS（1）?is?set?to?0?then?only?warning?messages?are?displayed.??If
%	OPTIONS（1）?is?-1?then?nothing?is?displayed.
%
%	OPTIONS（2）?is?a?measure?of?the?precision?required?for?the?value?of
%	the?weights?W?at?the?solution.
%
%	OPTIONS（3）?is?a?measure?of?the?precision?required?of?the?objective
%	function?at?the?solution.??Both?this?and?the?previous?condition?must
%	be?satisfied?for?termination.
%
%	OPTIONS（5）?is?set?to?1?if?an?approximation?to?the?Hessian?（which
%	assumes?that?all?outputs?are?independent）?is?used?for?softmax
%	outputs.?With?the?default?value?of?0?the?exact?Hessian?（which?is?more
%	expensive?to?compute）?is?used.
%
%	OPTIONS（14）?is?the?maximum?number?of?iterations?for?the?IRLS
%	algorithm;??default?100.
%
%	See?also
%	GLM?GLMERR?GLMGRAD
%

%	Copyright?（c）?Ian?T?Nabney?（1996-2001）

%?Check?arguments?for?consistency
errstring?=?consist（net?‘glm‘?x?t）;
if?~errstring
??error（errstring）;
end

if（~options（14））
??options（14）?=?100;
end

display?=?options（1）;
%?Do?we?need?to?test?for?termination?
test?=?（options（2）?|?options（3））;

ndata?=?size（x?1）;
%?Add?a?column?of?ones?for?the?bias?
inputs?=?[x?ones（ndata?1）];

%?Linear?outputs?are?a?special?case?as?they?can?be?found?in?one?step
if?strcmp（net.outfn?‘linear‘）
??if?~isfield（net?‘alpha‘）
????%?Solve?for?the?weights?and?biases?using?left?matrix?divide
????temp?=?inputs\t;
??elseif?size（net.alpha?==?[1?1]）
????%?Use?normal?form?equation
????hessian?=?inputs‘*inputs?+?net.alpha*eye（net.nin+1）;
????temp?=?pinv（hessian）*（inputs‘*t）;??
??else
????error（‘Only?scalar?alpha?allowed‘）;
??end
??net.w1?=?temp（1:net.nin?:）;
??net.b1?=?temp（net.nin+1?:）;
??%?Store?error?value?in?options?vector
??options（8）?=?glmerr（net?x?t）;
??return;
end

%?Otherwise?need?to?use?iterative?reweighted?least?squares
e?=?ones（1?net.nin+1）;
for?n?=?1:options（14）

??switch?net.outfn
????case?‘lo