You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@mahout.apache.org by Ted Dunning <te...@gmail.com> on 2014/09/01 00:13:32 UTC

Re: Urgent questions

See answers in-line



On Sun, Aug 31, 2014 at 4:17 AM, sara khodeir <sa...@gmail.com> wrote:

> I'm curious what the difference between the predictors and features are in
> the logistic regression model, what is the point of having features if it
> predicts using the predictors.
>

These are synonyms.  The reason for having two words is that features looks
more at the data while the word predictors looks more at what the feature
is doing in the model and is making a contrast with the phrase "target
variable".

Supervised learning has a sad abundance of redundant terminology.  Levels,
categories, models and so on come from different communities building the
same functionality independently with their own slant on how it should be
communicated.


> Also,I am trying to include more features than can fit into the command
> line so I wrote a file of the parameters and I'm trying to run it like this
> ./mahout "filepath" but this doesn't work. Does anyone know how I can do
> this?
>

How many features are you talking about having?  Also, which parts of
mahout and which version are you trying to use?

Re: Urgent questions

Posted by sara khodeir <sa...@gmail.com>.
Windows, they have a limit on the size of the command.


On Mon, Sep 1, 2014 at 10:06 AM, Ted Dunning <te...@gmail.com> wrote:

> What is your actually command line?
>
>
>
>
> On Sun, Aug 31, 2014 at 8:57 PM, sara khodeir <sa...@gmail.com>
> wrote:
>
> > About 20,000, I meant predictors though not features,  I'm using logistic
> > regression, as for the version, I 'm not sure it was pre-installed on a
> > cluster I'm using.
> >
> >
> > On Mon, Sep 1, 2014 at 1:13 AM, Ted Dunning <te...@gmail.com>
> wrote:
> >
> > > See answers in-line
> > >
> > >
> > >
> > > On Sun, Aug 31, 2014 at 4:17 AM, sara khodeir <sa...@gmail.com>
> > > wrote:
> > >
> > > > I'm curious what the difference between the predictors and features
> are
> > > in
> > > > the logistic regression model, what is the point of having features
> if
> > it
> > > > predicts using the predictors.
> > > >
> > >
> > > These are synonyms.  The reason for having two words is that features
> > looks
> > > more at the data while the word predictors looks more at what the
> feature
> > > is doing in the model and is making a contrast with the phrase "target
> > > variable".
> > >
> > > Supervised learning has a sad abundance of redundant terminology.
> > Levels,
> > > categories, models and so on come from different communities building
> the
> > > same functionality independently with their own slant on how it should
> be
> > > communicated.
> > >
> > >
> > > > Also,I am trying to include more features than can fit into the
> command
> > > > line so I wrote a file of the parameters and I'm trying to run it
> like
> > > this
> > > > ./mahout "filepath" but this doesn't work. Does anyone know how I can
> > do
> > > > this?
> > > >
> > >
> > > How many features are you talking about having?  Also, which parts of
> > > mahout and which version are you trying to use?
> > >
> >
>

Re: Urgent questions

Posted by Ted Dunning <te...@gmail.com>.
What is your actually command line?




On Sun, Aug 31, 2014 at 8:57 PM, sara khodeir <sa...@gmail.com> wrote:

> About 20,000, I meant predictors though not features,  I'm using logistic
> regression, as for the version, I 'm not sure it was pre-installed on a
> cluster I'm using.
>
>
> On Mon, Sep 1, 2014 at 1:13 AM, Ted Dunning <te...@gmail.com> wrote:
>
> > See answers in-line
> >
> >
> >
> > On Sun, Aug 31, 2014 at 4:17 AM, sara khodeir <sa...@gmail.com>
> > wrote:
> >
> > > I'm curious what the difference between the predictors and features are
> > in
> > > the logistic regression model, what is the point of having features if
> it
> > > predicts using the predictors.
> > >
> >
> > These are synonyms.  The reason for having two words is that features
> looks
> > more at the data while the word predictors looks more at what the feature
> > is doing in the model and is making a contrast with the phrase "target
> > variable".
> >
> > Supervised learning has a sad abundance of redundant terminology.
> Levels,
> > categories, models and so on come from different communities building the
> > same functionality independently with their own slant on how it should be
> > communicated.
> >
> >
> > > Also,I am trying to include more features than can fit into the command
> > > line so I wrote a file of the parameters and I'm trying to run it like
> > this
> > > ./mahout "filepath" but this doesn't work. Does anyone know how I can
> do
> > > this?
> > >
> >
> > How many features are you talking about having?  Also, which parts of
> > mahout and which version are you trying to use?
> >
>

Re: Urgent questions

Posted by sara khodeir <sa...@gmail.com>.
About 20,000, I meant predictors though not features,  I'm using logistic
regression, as for the version, I 'm not sure it was pre-installed on a
cluster I'm using.


On Mon, Sep 1, 2014 at 1:13 AM, Ted Dunning <te...@gmail.com> wrote:

> See answers in-line
>
>
>
> On Sun, Aug 31, 2014 at 4:17 AM, sara khodeir <sa...@gmail.com>
> wrote:
>
> > I'm curious what the difference between the predictors and features are
> in
> > the logistic regression model, what is the point of having features if it
> > predicts using the predictors.
> >
>
> These are synonyms.  The reason for having two words is that features looks
> more at the data while the word predictors looks more at what the feature
> is doing in the model and is making a contrast with the phrase "target
> variable".
>
> Supervised learning has a sad abundance of redundant terminology.  Levels,
> categories, models and so on come from different communities building the
> same functionality independently with their own slant on how it should be
> communicated.
>
>
> > Also,I am trying to include more features than can fit into the command
> > line so I wrote a file of the parameters and I'm trying to run it like
> this
> > ./mahout "filepath" but this doesn't work. Does anyone know how I can do
> > this?
> >
>
> How many features are you talking about having?  Also, which parts of
> mahout and which version are you trying to use?
>