Machine Learning - Revision history

Leo: Text replacement - "{\| " to "{| class="wikitable" "

2019-09-01T01:33:00Z

Text replacement - "{\| " to "{| class="wikitable" "

← Older revision		Revision as of 01:33, 1 September 2019
Line 105:		Line 105:

	== Glossary ==		== Glossary ==
	{\|		{\| class="wikitable"
	! Word		! Word
	! Definition		! Definition

Leo: /* Tensorflow */

2017-10-14T22:02:38Z

Tensorflow

← Older revision		Revision as of 22:02, 14 October 2017
Line 136:		Line 136:
	All operations on the graph are executed in a Tensorflow session. A session is an object that runs operations on the graph and tracks the state of each node in the graph.		All operations on the graph are executed in a Tensorflow session. A session is an object that runs operations on the graph and tracks the state of each node in the graph.

			=== Installation ===
			Simplest way to install TensorFlow is to use pip:
			{{highlight\|lang=terminal\|code=
			# pip3 install tensorflow
			}}

			For GPU support, use the {{code\|tensorflow-gpu}} package.

			=== Introduction ===

			Dealing with nodes. Values that get changed every iteration should be placeholder nodes.

			{{highlight\|lang=python\|code=
			import tensorflow as tf

			# Model parameters
			# variable: default value, type
			W = tf.Variable([.3], dtype=tf.float32)
			b = tf.Variable([-.3], dtype=tf.float32)

			# Placeholders are variables that are populated with training data
			# Input training data, and output 'answers' to check the loss value
			x = tf.placeholder(tf.float32)
			y = tf.placeholder(tf.float32)

			# The model we are training with. We want to solve for W and b.
			linear_model = W * x + b

			# loss function.
			# Sum Squares; squares to penalize larger differences
			loss = tf.reduce_sum(tf.square(linear_model - y))

			# optimizer function to 'adjust' values
			optimizer = tf.train.GradientDescentOptimizer(0.01)
			# We want to minimize loss
			train = optimizer.minimize(loss)

			# training data
			x_train = [1, 2, 3, 4, 5]
			y_train = [0, -1, -2, -3, -4 ]

			# training loop
			init = tf.global_variables_initializer()
			sess = tf.Session()
			sess.run(init) # reset values to wrong

			# Training N iterations
			for i in range(1000):
			sess.run(train, {x: x_train, y: y_train})

			# evaluate training accuracy
			curr_W, curr_b, curr_loss = sess.run([W, b, loss], {x: x_train, y: y_train})
			print("W: %s b: %s loss: %s"%(curr_W, curr_b, curr_loss))

			}}

	== Other Topics ==		== Other Topics ==

Leo at 02:42, 3 September 2017

2017-09-03T02:42:53Z

@@ Line 124: / Line 124: @@
 | Another name for value prediction.
 |}
 == Other Topics ==

Leo: /* Glossary */

2017-09-03T01:02:06Z

Glossary

← Older revision		Revision as of 01:02, 3 September 2017
Line 110:		Line 110:
	\|-		\|-
	\| Overfitting		\| Overfitting
	\| Where the model is too specific to the training data.		\| Where the model is too specific to the training data and 'memorizes' the exact training data; it does not figure out the pattern in the data. You can see this if training set error is low, but test data error is high.
			\|-
			\| Underfitting
			\| Where the model is too simple and doesn't fully learn the patterns in the data. You can see this if both training set and test set errors are very high.
	\|-		\|-
	\| Features		\| Features

Leo: /* Glossary */

2017-09-03T00:47:06Z

Glossary

← Older revision		Revision as of 00:47, 3 September 2017
Line 117:		Line 117:
	\| Value to Predict		\| Value to Predict
	\| The value to predict is the value generated from a learning model. Called Y in machine learning code.		\| The value to predict is the value generated from a learning model. Called Y in machine learning code.
			\|-
			\| Regression
			\| Another name for value prediction.
	\|}		\|}

Leo: /* Hands On Overview */

2017-09-03T00:39:38Z

Hands On Overview

@@ Line 88: / Line 88: @@
 # Create a tree that's hundreds of layers deep with thousands of branching paths or
 # Create many small trees that contribute a little to the final answer (called ensemble learning)
 == Glossary ==

Leo: /* Glossary */

2017-09-03T00:26:30Z

Glossary

← Older revision		Revision as of 00:26, 3 September 2017
Line 98:		Line 98:
	\|-		\|-
	\| Features		\| Features
	\| The fields in a piece of data that can be used to determine the answer		\| The fields in a piece of data that can be used to determine the answer. Called X in machine learning code.
	\|-		\|-
	\| ~~Overfitting~~		\| Value to Predict
	\| ~~Where~~ the model ~~is too specific to the training data~~.		\| The value to predict is the value generated from a learning model. Called Y in machine learning code.
	\|}		\|}


	== Other Topics ==		== Other Topics ==

Leo at 00:24, 3 September 2017

2017-09-03T00:24:58Z

← Older revision		Revision as of 00:24, 3 September 2017
Line 4:		Line 4:

	Supervised learning requires training a model with labeled training data (data with parameters and a solution). The supervised learning algorithm then generates a solution that is as close to the solution as possible. This solution can then be applied to any other problem of the same type.		Supervised learning requires training a model with labeled training data (data with parameters and a solution). The supervised learning algorithm then generates a solution that is as close to the solution as possible. This solution can then be applied to any other problem of the same type.

			{{highlight\|lang=text\|code=
			[ Features ] ----> [ Supervised Learning Model ] ----> [ Prediction ]
			}}

			How the supervised learning model works depends on the implementation. A multivariate linear regression approach will give each feature a weight and iteratively determine a set of weights that produces a prediction that is as close to the answer as possible.


	Unsupervised learning is basically what data mining is: it attempts to look for hidden patterns in the data.		Unsupervised learning is basically what data mining is: it attempts to look for hidden patterns in the data.


	=== Multivariate Linear Regression ===		=== Multivariate Linear Regression ===
Line 85:		Line 93:
	! Word		! Word
	! Definition		! Definition
			\|-
			\| Overfitting
			\| Where the model is too specific to the training data.
			\|-
			\| Features
			\| The fields in a piece of data that can be used to determine the answer
	\|-		\|-
	\| Overfitting		\| Overfitting

Leo: /* Hands On Overview */

2017-09-03T00:21:36Z

Hands On Overview

@@ Line 41: / Line 41: @@
 sizes = sizes * 0.3
 }}
@@ Line 56: / Line 80: @@
 # Create a tree that's hundreds of layers deep with thousands of branching paths or
 # Create many small trees that contribute a little to the final answer (called ensemble learning)
 == Glossary ==

Leo at 00:17, 3 September 2017

2017-09-03T00:17:05Z

@@ Line 20: / Line 20: @@
 By squaring each difference, we penalize larger errors more.
-Step 3 requires picking new weights that reduces the cost. This can be done by something called a batch gradient descent. A gradient descent is an iterative optimization algorithm that can be used to minimize the cost function and to find the best weights. However, the data that can be handled with this approach has to be linear.
+Step 3 requires picking new weights that reduces the cost. This can be done by something called a batch gradient descent. A gradient descent is an iterative optimization algorithm that can be used to minimize the cost function and to find the best weights. However, the data that can be handled with this approach has to be linear. You can use other algorithms such as gradient boosting for non-linear data.
 == Other Topics ==