TODO: Lienar Regression with least Mean Squares (LMS) Optimize the model through gradient descent. *Please complete the TODOs. * !pip install wget import os import random import traceback from pdb import set_trace import sys import numpy as np from abc import ABC, abstractmethod import traceback from util.timer import Timer from util.data import split_data, feature_label_split, Standardization from util.metrics import mse from datasets.HousingDataset import HousingDataset class BaseModel(ABC): """ Super class for ITCS Machine Learning Class""" @abstractmethod def fit(self, X, y): pass @abstractmethod def predict(self, X): pass class LinearModel(BaseModel): """ Abstract class for a linear model Attributes ========== w ndarray weight vector/matrix """ def __init__(self): """ weight vector w is initialized as None """ self.w = None # check if the matrix is 2-dimensional. if not, raise an exception def _check_matrix(self, mat, name): if len(mat.shape) != 2: raise ValueError(f"Your matrix {name} shape is not 2D! Matrix {name} has the shape {mat.shape}") # add a biases def add_ones(self, X): """ add a column basis to X input matrix """ self._check_matrix(X, 'X') return np.hstack((np.ones((X.shape[0], 1)), X)) #################################################### #### abstract funcitons ############################ @abstractmethod def fit(self, X: np.ndarray, y: np.ndarray): """ train linear model Args: X: Input data y: targets/labels """ pass @abstractmethod def predict(self, X: np.ndarray): """ apply the learned model to input X parameters ---------- X 2d array input data """ pass class LeastMeanSquares(LinearModel): """ Performs regression using least mean squares (gradient descent) attributes: w (np.ndarray): weight matrix alpha (float): learning rate or step size epochs (int): Number of epochs to run for mini-batch gradient descent seed (int): Seed to be used for NumPy's RandomState class or universal seed np.random.seed() function. """ def __init__(self, alpha: float, epochs: int, seed: int = None): super().__init__() self.w = None self.alpha = alpha self.epochs = epochs self.seed = seed def fit(self, X: np.ndarray, y: np.ndarray) -> None: """ Used to train our model to learn optimal weights. TODO: Finish this method by adding code to perform LMS in order to learn the weights `self.w`. """ pass # TODO replace this line with your code def predict(self, X: np.ndarray) -> np.ndarray: """ Used to make a prediction using the learned weights. TODO: Finish this method by adding code to make a prediction given the learned weights `self.w`. """ # TODO (REQUIRED) Add code below # TODO (REQUIRED) Store predictions below by replacing np.ones() y_hat = np.ones([len(X), 1]) return y_hat
TODO: Lienar Regression with least Mean Squares (LMS)
Optimize the model through gradient descent. *Please complete the TODOs. *
!pip install wget
import os
import random
import traceback
from pdb import set_trace
import sys
import numpy as np
from abc import ABC, abstractmethod
import traceback
from util.timer import Timer
from util.data import split_data, feature_label_split, Standardization
from util.metrics import mse
from datasets.HousingDataset import HousingDataset
class BaseModel(ABC):
""" Super class for ITCS Machine Learning Class"""
@abstractmethod
def fit(self, X, y):
pass
@abstractmethod
def predict(self, X):
pass
class LinearModel(BaseModel):
"""
Abstract class for a linear model
Attributes
==========
w ndarray
weight
"""
def __init__(self):
"""
weight vector w is initialized as None
"""
self.w = None
# check if the matrix is 2-dimensional. if not, raise an exception
def _check_matrix(self, mat, name):
if len(mat.shape) != 2:
raise ValueError(f"Your matrix {name} shape is not 2D! Matrix {name} has the shape {mat.shape}")
# add a biases
def add_ones(self, X):
"""
add a column basis to X input matrix
"""
self._check_matrix(X, 'X')
return np.hstack((np.ones((X.shape[0], 1)), X))
####################################################
#### abstract funcitons ############################
@abstractmethod
def fit(self, X: np.ndarray, y: np.ndarray):
"""
train linear model
Args:
X: Input data
y: targets/labels
"""
pass
@abstractmethod
def predict(self, X: np.ndarray):
"""
apply the learned model to input X
parameters
----------
X 2d array
input data
"""
pass
class LeastMeanSquares(LinearModel):
"""
Performs regression using least mean squares (gradient descent)
attributes:
w (np.ndarray): weight matrix
alpha (float): learning rate or step size
epochs (int): Number of epochs to run for mini-batch
gradient descent
seed (int): Seed to be used for NumPy's RandomState class
or universal seed np.random.seed() function.
"""
def __init__(self, alpha: float, epochs: int, seed: int = None):
super().__init__()
self.w = None
self.alpha = alpha
self.epochs = epochs
self.seed = seed
def fit(self, X: np.ndarray, y: np.ndarray) -> None:
""" Used to train our model to learn optimal weights.
TODO:
Finish this method by adding code to perform LMS in order to learn the
weights `self.w`.
"""
pass # TODO replace this line with your code
def predict(self, X: np.ndarray) -> np.ndarray:
""" Used to make a prediction using the learned weights.
TODO:
Finish this method by adding code to make a prediction given the learned
weights `self.w`.
"""
# TODO (REQUIRED) Add code below
# TODO (REQUIRED) Store predictions below by replacing np.ones()
y_hat = np.ones([len(X), 1])
return y_hat
Step by step
Solved in 2 steps
