ARIMA

ARIMA (AutoRegressive Integrated Moving Average) is a popular time series forecasting model that combines autoregressive (AR), differencing (I), and moving average (MA) components to capture the patterns and dependencies in time series data. It is widely used for analyzing and forecasting time-dependent data points.

The ARIMA model is suitable for stationary time series data, where the statistical properties such as mean and variance remain constant over time. If the data is non-stationary, differencing can be applied to make it stationary by removing trends or seasonality.

Method: POST Authorization: API Key

https://engine.raccoon-ai.io/api/v1/ml/time-series/arima

Authorization

Type	Key	Value
API Key	X-Api-Key	rae_######

Request Body

Section	Key	Data Type	Required	Description
train	data	json	true	Data that use to train the model
	date_col	string	true	Input features (X)
	target_col	string	true	Output targets (y)
config	freq	string	false	Gap between datas/ time
	test_size	float	false	Test size for split data
forcast	forcast_for	int	true	Number of points that need to forcast

Types

{
  "train": {
    "data": <json>,
    "dates_col": <string>,
    "target_col": <string>
  },
  "config": {
    "freq": <string>,
    "test_size": <float>
  },
  "forcast_for": <int>
}

Sample

{
  "train": {
    "data": {
      "dates": {
        "0": "2022-11-25",
        "1": "2022-12-02",
        "2": "2022-12-09",
        "3": "2022-12-16",
        "4": "2022-12-23",
        "5": "2022-12-30",
        "6": "2023-01-06",
        "7": "2023-01-13",
        "8": "2023-01-20",
        "9": "2023-01-27"
      },
      "marks": {
        "0": 161,
        "1": 123,
        "2": 134,
        "3": 167,
        "4": 143,
        "5": 156,
        "6": 167,
        "7": 143,
        "8": 156,
        "9": 167
      }
    },
    "dates_col": "dates",
    "target_col": "marks"
  },
  "config": {
    "freq": "W",
    "test_size": 0.25
  },
  "forcast_for": 5
}

Reponse Body

Key	Data Type	Description
success	boolean	Indicate the success of the request
msg	string	Message indicators
error	string	Error information, only set if success is false
result	json	Result, only set if success is true
score	json	RMSE (Root Mean Square Error) of the model, only set if success is true
generated_ts	float	Generated timestamp

Types

{
  "success": <boolean>,
  "msg": <string | null>,
  "error": <string | null>,
  "result": <list>,
  "score": {
    "rmse": <float>
  },
  "generated_ts": <timestamp>
}

Sample

{
  "success": true,
  "msg": "Model trained and predicted successfully",
  "error": null,
  "result": [
    145.306460384454, 159.01596373700463, 165.26655276778865,
    144.69950483108218, 160.2060306291415
  ],
  "score": {
    "rmse": 14.027130973766175
  },
  "generated_ts": 1685514898.064395
}