Skip to content

Latest commit

 

History

History

cntk-lstm-forecast

README.md

singnetlogo

CNTK Time Series Forecast

This service uses CNTK Time Series Prediction with LSTM and Python SAX to extrapolate a given time series.

It is part of our Time Series Analysis Services.

Getting Started

Requirements

Development

Clone this repository:

$ git clone https://github.com/singnet/time-series-analysis.git
$ cd generic/cntk-lstm-forecast

Running the service:

To get the ORGANIZATION_ID and SERVICE_ID you must have already published a service (check this link).

Create the SNET Daemon's config JSON file (snetd.config.json).

{
   "DAEMON_END_POINT": "DAEMON_HOST:DAEMON_PORT",
   "IPFS_END_POINT": "http://ipfs.singularitynet.io:80",
   "BLOCKCHAIN_NETWORK_SELECTED": "BLOCKCHAIN_NETWORK",
   "PASSTHROUGH_ENDPOINT": "http://SERVICE_GRPC_HOST:SERVICE_GRPC_PORT",  
   "ORGANIZATION_ID": "ORGANIZATION_ID",
   "SERVICE_ID": "SERVICE_ID",
   "LOG": {
       "LEVEL": "debug",
       "OUTPUT": {
            "TYPE": "stdout"
           }
   }
}

For example (using the Ropsten testnet):

$ cat snetd.config.json
{
   "DAEMON_END_POINT": "0.0.0.0:7059",
   "IPFS_END_POINT": "http://ipfs.singularitynet.io:80",
   "BLOCKCHAIN_NETWORK_SELECTED": "ropsten",
   "PASSTHROUGH_ENDPOINT": "http://localhost:7003",
   "ORGANIZATION_ID": "snet",
   "SERVICE_ID": "cntk-lstm-forecast",
   "LOG": {
       "LEVEL": "debug",
       "OUTPUT": {
           "TYPE": "stdout"
           }
   }
}

Note that we set DAEMON_HOST = 0.0.0.0 because this service will run inside a Docker container.

Install all dependencies:

$ pip3 install -r requirements.txt

Generate the gRPC codes:

$ sh buildproto.sh

Start the service and SNET Daemon:

$ python3 run_service.py

Calling the service:

Inputs:

  • window_len: the SAX window length.
  • word_len: the SAX word length.
  • alphabet_size: the SAX alphabet size.
  • source_type: from CSV file ("csv") or financial data ("financial").
    • For "csv":
      • The CSV file must have a column with name "input".
      • The data length that service supports is (CSV rows) * window_len <= 200k`.
  • source:
    • If source_type: "csv", a URL to CSV file.
    • If source_type: "financial", the source to get market data (ie. yahoo, check this link).
  • contract:
    • If source_type: "csv", empty.
    • If source_type: "financial", label of asset (like "SPY", check this link).
  • start_date:
    • If source_type: "csv", empty.
    • If source_type: "financial", start date of training dataset (format "YYYY-MM-DD").
  • end_date:
    • If source_type: "csv", empty.
    • If source_type: "financial", end date of training dataset (format "YYYY-MM-DD").

Note: The date delta must be >= window_len.

Local (testing purpose):

$ python3 test_service.py
Endpoint (localhost:7003): 
Method (forecast): 
Window length (24): 
Word length (8): 
Alphabet length (5): 
Source Type (financial): 
Source (yahoo): 
Contract (SPY): 
Start date (2012-01-01): 
End date (2018-12-10):

response:
last_sax_word           : "deccabeb"
forecast_sax_letter     : "c"
position_in_sax_interval: 0.14
  • last_sax_word: The last SAX's word of the series.
  • forecast_sax_letter: a SAX's letter that model has forecast.
  • position_in_sax_interval: is a number in [0, 1]. It's a measure of how much the prediction is inside a SAX's letter interval.
    • "c" and position_in_sax_interval tending to 0.00 (0.14) means that the prediction is at the beginning of SAX letter's interval.

for further instructions about the output of this service, check the User's Guide.

Through SingularityNET (follow this link to learn how to publish a service and open a payment channel to be able to call it):

Assuming that you have an open channel (id: 0) to this service:

$ snet client call cntk-lstm-forecast forecast '{"window_len": 36, "word_len": 18, "alphabet_size": 5, "source_type": "financial", "source": "yahoo", "contract": "AAPL", "start_date": "2017-01-01", "end_date": "2018-12-10"}'

last_sax_word: "eeeeeedddcbbaaaaaa"
forecast_sax_letter: "c"
position_in_sax_interval: 0.2819729149341583

Contributing and Reporting Issues

Please read our guidelines before submitting an issue. If your issue is a bug, please use the bug template pre-populated here. For feature requests and queries you can use this template.

Authors