prediction.md

description
Predict the time series data

Prediction

Give a data frame with these columns

Date	Close
2014-04-14	133.95

Base data can come from Yahoo driver

All

The model will predict SVR, LINEAR, ARIMA for 20 days on Close value

dataset = naas_drivers.yahoo.stock("TSLA")
pr = naas_drivers.prediction.get(dataset=dataset, prediction_type="all")

Prediction size

• data_points: The number of days in the future that are to predict.

dataset = naas_drivers.yahoo.stock("TSLA")
pr = naas_drivers.prediction.get(dataset=dataset, data_points=50)

Model

All the parameters of the above formula are explained below.

• prediction_type: The model to predict, it can be SVR, LINEAR, ARIMA, or all

dataset = naas_drivers.yahoo.stock("TSLA")
pr = naas_drivers.prediction.get(dataset=dataset, prediction_type="ARIMA")

Options

• dataset : the dataset in DataFrame format
• column: The exact name of the column that is to be predicted, from the dataset
• date_column:The date range from the dataset. Will be used as the output index.
• prediction_type: Can be ARIMA, LINEAR, SVR, COMPOUND or all
• data_points **(optional):** number of days to predict
• concact_label (optional): A column name who will generate a concatenated frame with past and future data.

dataset = naas_drivers.yahoo.stock("TSLA")
pr = naas_drivers.prediction.get(dataset=dataset)

Plot

Once you have predicted using the above predict formula, you can plot the predictions

  naas_drivers.plotly.stock(pr, , "linechart_close")

Check more options on the link below