Hi Weiru, Remove the `clipInput` entries from the following sections in params_coordinate_pre.py:
temperature vector _classifierInput timestamp_dayOfWeek - Chetan > On Jan 17, 2016, at 6:44 PM, Weiru Zeng <[email protected]> wrote: > > Hi Chetan: > > I'm sorry,I send a wrong e-mail just now. > > I study the code in the https://github.com/numenta/nupic.geospatial > <https://github.com/numenta/nupic.geospatial>. And I want to add the > temperature, refer to the > https://github.com/numenta/nupic/tree/master/examples/opf/clients/hotgym/anomaly/one_gym, > I writed the code attached in the e-mail. This code is make a > TemporalMultiStep type inference, and the input is timestamp, coordinate, > temperature, I want to predict the coordinate and make anormally score for > the coordinate. Running the code, there is a error in the > params_coordinate_pre.py like: > > TypeError: __init__() got an unexpected keyword argument 'clipInput'. > > > I check the params_coordinate_pre.py for a long time, but can't solve the > problem, so could you please help me. > > The parameters is hard to set without the swarm : ( . Hope for a better swarm. > > Thanks for your help! > > Weiru Zeng > > ------------------ Original ------------------ > From: "Chetan Surpur";<[email protected]>; > Date: Tue, Jan 12, 2016 03:53 AM > To: "Weiru Zeng"<[email protected]>; > Subject: Re: Help for Running warms with coodinate included .csv file > > Hi Weiru, > > Unfortunately, it is not possible at this time to swarm over > CoordinateEncoder parameters. So far we've been manually creating the > parameters when we want to use CoordinateEncoder. > > It looks like you're trying to predict the coordinate (since you said "the > inference field is coordinate"). Do you think temperature would help with > that, or is it possible to predict the next coordinate given just the history > of coordinates so far? If so, then you can start with the model parameters in > the nupic.geospatial example [1], and tweak for your application. > > [1] https://github.com/numenta/nupic.geospatial > <https://github.com/numenta/nupic.geospatial> > > - Chetan > >> On Jan 9, 2016, at 2:02 AM, Weiru Zeng <[email protected] >> <mailto:[email protected]>> wrote: >> >> Hi everyone: >> >> I have some data which contain the coordinate, and I want to use the swarms >> to get a good model_params.py file. But the type used to feed the swarms is >> not include the coodinate type. It just contains the type below: datetime, >> int, float, bool, string, list, for the list type, the official document >> said that it is third-party - not support. >> >> I have received a model_params.py which is configured artficial from a >> friend. But the parameters in the file is rough, it's hard to adapt the >> requirment. >> >> So I want to know how can I use the swarms handle the coordinate type data >> to make the opf use the coordinateEncoder to encode. or what's you way to >> deal with the coordinate included file by the opf. I want to handle the >> x,y,z as the coordinate, but not three independent number. >> >> My data is like: >> id,time,x,y,z,temperature >> 38,2015/11/5 18:14:59,184,-184,-5,66 >> 38,2015/11/5 18:15:00,184,-184,-5,66 >> 38,2015/11/5 18:15:01,184,-184,-6,66 >> 38,2015/11/5 18:15:01,184,-184,-6,66 >> 38,2015/11/5 18:15:01,184,-184,-5,66 >> 38,2015/11/5 18:15:03,184,-184,-6,66 >> 38,2015/11/5 18:15:03,184,-184,-6,66 >> 38,2015/11/5 18:15:04,184,-184,-5,66 >> 38,2015/11/5 18:15:05,184,-184,-6,66 >> 38,2015/11/5 18:15:06,184,-184,-5,66 >> 38,2015/11/5 18:15:06,184,-184,-6,66 >> 38,2015/11/5 18:15:07,184,-184,-5,66 >> >> The input columns is time, coordinate(x,y,z), temperature. The >> inferencefield is coordinate. >> >> Thanks! >> >> Weiru Zeng >
