doubt.datasets.servo

  1"""Servo data set.
  2
  3This data set is from the UCI data set archive, with the description being the original
  4description verbatim. Some feature names may have been altered, based on the
  5description.
  6"""
  7
  8import io
  9
 10import pandas as pd
 11
 12from .dataset import BASE_DATASET_DESCRIPTION, BaseDataset
 13
 14
 15class Servo(BaseDataset):
 16    __doc__ = f"""
 17    Data was from a simulation of a servo system.
 18
 19    Ross Quinlan:
 20
 21    This data was given to me by Karl Ulrich at MIT in 1986. I didn't record his
 22    description at the time, but here's his subsequent (1992) recollection:
 23
 24    "I seem to remember that the data was from a simulation of a servo system involving
 25    a servo amplifier, a motor, a lead screw/nut, and a sliding carriage of some sort.
 26    It may have been on of the translational axes of a robot on the 9th floor of the AI
 27    lab. In any case, the output value is almost certainly a rise time, or the time
 28    required for the system to respond to a step change in a position set point."
 29
 30    (Quinlan, ML'93)
 31
 32    "This is an interesting collection of data provided by Karl Ulrich. It covers an
 33    extremely non-linear phenomenon - predicting the rise time of a servomechanism in
 34    terms of two (continuous) gain settings and two (discrete) choices of mechanical
 35    linkages."
 36
 37    {BASE_DATASET_DESCRIPTION}
 38
 39    Features:
 40        motor (int):
 41            Motor, ranges from 0 to 4 inclusive
 42        screw (int):
 43            Screw, ranges from 0 to 4 inclusive
 44        pgain (int):
 45            PGain, ranges from 3 to 6 inclusive
 46        vgain (int):
 47            VGain, ranges from 1 to 5 inclusive
 48
 49    Targets:
 50        class (float):
 51            Class values, ranges from 0.13 to 7.10 inclusive
 52
 53    Source:
 54        https://archive.ics.uci.edu/ml/datasets/Servo
 55
 56    Examples:
 57        Load in the data set::
 58
 59            >>> dataset = Servo()
 60            >>> dataset.shape
 61            (167, 5)
 62
 63        Split the data set into features and targets, as NumPy arrays::
 64
 65            >>> X, y = dataset.split()
 66            >>> X.shape, y.shape
 67            ((167, 4), (167,))
 68
 69        Perform a train/test split, also outputting NumPy arrays::
 70
 71            >>> train_test_split = dataset.split(test_size=0.2, random_seed=42)
 72            >>> X_train, X_test, y_train, y_test = train_test_split
 73            >>> X_train.shape, y_train.shape, X_test.shape, y_test.shape
 74            ((131, 4), (131,), (36, 4), (36,))
 75
 76        Output the underlying Pandas DataFrame::
 77
 78            >>> df = dataset.to_pandas()
 79            >>> type(df)
 80            <class 'pandas.core.frame.DataFrame'>
 81    """
 82
 83    _url = (
 84        "https://archive.ics.uci.edu/ml/machine-learning-databases/" "servo/servo.data"
 85    )
 86
 87    _features = range(4)
 88    _targets = [4]
 89
 90    def _prep_data(self, data: bytes) -> pd.DataFrame:
 91        """Prepare the data set.
 92
 93        Args:
 94            data (bytes): The raw data
 95
 96        Returns:
 97            Pandas dataframe: The prepared data
 98        """
 99        # Convert the bytes into a file-like object
100        csv_file = io.BytesIO(data)
101
102        # Load in the dataframe
103        cols = ["motor", "screw", "pgain", "vgain", "class"]
104        df = pd.read_csv(csv_file, names=cols)
105
106        # Encode motor and screw
107        codes = ["A", "B", "C", "D", "E"]
108        df["motor"] = df.motor.map(lambda x: codes.index(x))
109        df["screw"] = df.screw.map(lambda x: codes.index(x))
110
111        return df