Outputs

This section summarizes the mathematical outputs of TIM Detect with an outlier approach. Note that this approach does not return accuracies in the outputs because there is no target or label to refer to.

CSV result (table)

build-model, detect.

column_name	timestamp	anomlay_indicator	outlier
Temperature	2020-10-12T03:00:00.0	0.31	false
Temperature	2020-10-12T04:00:00.0	0.45	false
Temperature	2020-10-12T05:00:00.0	1.27	true
Temperature	2020-10-12T06:00:00.0	1.23	true
Pressure	2020-10-12T03:00:00.0	0.56	false
Pressure	2020-10-12T04:00:00.0	1.13	true
Pressure	2020-10-12T05:00:00.0	0.82	false
Pressure	2020-10-12T06:00:00.0	0.33	false

Column name

The column_name represents the name of the column from dataset on which was the outlier detection performed.

Timestamp

The timestamp column represents the timestamp that corresponds to the given row of outputs.

Anomaly indicator

The anomlay_indicator column contains numbers from the interval (0, infinity) that specify the extent to which a given data point is outlier for the given column. Data points with an anomaly indicator higher than 1 are considered outliers. See the anomaly indicator section to learn more.

Outlier

The outlier column contains boolean values indicating whether the given point in time is an outlier for the given column.

API Model

Model version

The version of model. Each approach has its own version.

"modelVersion": "5.0"

Approach

The approach used to build the model.

"approach": "outlier"

Model

The model for outlier detection contains three parameters settings, models and parameters.

Settings

Stored settings which were used to build the model. There are three of them:

• rows - rows used to build model. Rows are always stored as an array of "from", "to" ranges even if they were entered relatively.
• sensitivity - sensitivity entered in the configuration
• maxModelComplexity - model complexity entered in the configuration

"settings": {
    "rows": [
        {
            "to": "2015-09-30T23:00:00.000Z",
            "from": "2014-10-01T00:00:00.000Z"
        }
    ],
    "sensitivity": 1,
    "maxModelComplexity": 30
}

Models

Array of gaussian mixture models for each variable. Contains four parameters:

• variable name - name of variable for which is the model built
• probability distribution - stored the GMM model
• threshold - constant used to compute the anomaly indicator
• translation - constant used to compute the anomaly indicator

"models": [
    {
        "variableName": "GasCons",
        "probabilityDistribution": {
            "n": 4,
            "d": 1,
            "w": [
                0.2249686741,
                0.386395561,
                0.3610753897,
                0.02756037521
            ],
            "μ": [
                90740.53879,
                65936.31899,
                82129.86789,
                38325.94086
            ],
            "Σ": [
                67477185.56,
                36296143.83,
                42240916.34,
                56937921.56
            ]
        },
        "threshold": 0.1015357793012146,
        "translation": 1.4158793011004753
    }
]

Parameters

Parameters describe the data properties for which the model was built. There are two of them:

• sampling period - the time scale given by user, if not given the estimated sampling period of the data. The sampling period used when building the model is stored in the ISO 8601 duration format.
• time zone - the time zone given by dataset

"parameters": {
    "samplingPeriod": "P1D",
    "timeZone": "UTC"
}

Signature

The signature serves to verify model originality.

"signature": "395b068eb6747efe4f9eb78"