Onnowpurbo: Created page with "Sumber: https://stackoverflow.com/questions/47408427/tensorflow-timeseries-data-import-with-datetime According to the source code, a RandomWindowInputFn accepts either a..."

2019-08-06T22:04:25Z

Created page with "Sumber: https://stackoverflow.com/questions/47408427/tensorflow-timeseries-data-import-with-datetime According to the source code, a RandomWindowInputFn accepts either a..."

New page

Sumber: https://stackoverflow.com/questions/47408427/tensorflow-timeseries-data-import-with-datetime

According to the source code, a RandomWindowInputFn accepts either a CSVReader or a NumpyReader. So you could use pandas to read the CSV, do the date parsing and then feed the transformed dates into a NumpyReader

My time-series data looks like this

timestamp value
0 2014-02-14 14:30:00 0.132
1 2014-02-14 14:35:00 0.134
2 2014-02-14 14:40:00 0.134
3 2014-02-14 14:45:00 0.134
4 2014-02-14 14:50:00 0.134

First, i parsed the timestamp column into a int col using pandas

from datetime import datetime as dt
import pandas as pd

def date_parser(date_str):
return dt.strptime(date_str, "%Y-%m-%d %H:%M:%S").strftime("%s")

data = pd.read_csv("my_data.csv"
, header=0
, parse_dates=['timestamp']
, date_parser=date_parser)

data['timestamp'] = data['timestamp'].apply(lambda x: int(x))

Then we can pass on these arrays to the NumpyReader

np_reader = tf.contrib.timeseries.NumpyReader(data={tf.contrib.timeseries.TrainEvalFeatures.TIMES: data['timestamp'].values, tf.contrib.timeseries.TrainEvalFeatures.VALUES : data['value'].values})

And finally pass the np_reader to the RandomWindowInputFn

train_input_fn = tf.contrib.timeseries.RandomWindowInputFn(
np_reader, batch_size=32, window_size=16)

Hope this helps somebody!

==Pranala Menarik==

* [[Keras]]

Keras: read csv timeseries - Revision history

Onnowpurbo: Created page with "Sumber: https://stackoverflow.com/questions/47408427/tensorflow-timeseries-data-import-with-datetime According to the source code, a RandomWindowInputFn accepts either a..."