Machine Learning - Coursera - Machine Learning
Specialization Questions With Complete Solutions
1st order Markov models Correct Answers Restricted to
encoding sequential correlation on previous element only
3 categories of machine learning algorithms Correct Answers 1.
Supervised learning which uses labeled data to predict new data
2. Unsupervised learning which finds patterns in unlabeled data
and
3. Reinforcement learning, which trains a model to make
decisions based on rewards and punishments from the
environment
A Bernoulli trial Correct Answers - It is a trial with a binary
outcome, for which the probability that the outcome is 1 equals
p (think of a coin toss of an old warped coin with the probability
of throwing heads being p).
- A Bernoulli experiment is a number of Bernoulli trials
performed after each other. These trials are i.i.d. by definition.
A Latice/Trellis diagram visualizes Correct Answers state
transitions over time
Also good tool to to visualize optimal path through states
(Viterbi Algorithm)
Abstract Essence of ML Correct Answers Representation +
Evaluation + Optimization
accuracy Correct Answers (TP + TN)/ (TP + TN + FP + FN)
,accuracy may not be useful measure in cases where Correct
Answers 1- There is a large class skew
2- There are differential misclassification costs - say, getting a
positive wrong costs more than getting a negative wrong.
3- We are interested in a subset of high confidence predictions
Acquiring emissions Correct Answers Wide range of options to
model ****** probabilities:
- Discrete tables
- Gaussians
- Mixture of Gaussians
- Neural Networks/RVMs etc to mode
All probability theory can be expressed in terms of two rules
Correct Answers - Product rule
- Sum rule
Ancestral sampling Correct Answers is a simple sampling
method well suited to PGNs
ANN feature selection Correct Answers • Artificial Neural
Networks can implicitly perform feature selection
• A multi-layer neural network where the first hidden layer has
fewer units (nodes) than the input layer
• Called 'Auto-associative' networks
Applications of machine learning Correct Answers Self Driving
Cars
Image Classification
Text Classification
Search Engine
,Banking, Healthcare Domain. web searches, photo recognition,
movie recommendations, and spam filtering. It is also being
implemented in industries, including optimizing wind turbines,
healthcare, and factory inspection.
Applications of supervised learning Correct Answers
Classification is a machine learning technique used for tasks like
spam filtering, online advertising, and visual inspection to
classify data into specific categories.
Regression is another machine learning technique used for tasks
like speech recognition, machine translation, and predicting the
position of objects in self-driving cars, where the goal is to
predict numerical values based on input data.
Apriori algorithm Correct Answers Apriori algorithm is a fast
way of finding frequent itemsets
ARFF Correct Answers Attribute-Relation File Format
Artificial Neural Nets Correct Answers • Feed-forward neural
network/Multilayer Perceptron one of many ANNs
• We focus on the Multilayer Perceptron
• Really multiple layers of logistic regression models
Association Rules Correct Answers Reflect items that are
frequently found (purchased) together, i.e. they are frequent
itemsets
• Information that customers who buy beer also buy crisps is e.g.
encoded as:
beer ) crisps[support = 2%, confidence = 75%]
, Attribute subset selection Correct Answers Feature selection
Feature selection is a form of dimensionality reduction in ML,
hence the DM term 'dimensionality reduction' for manifold
projection is problematic.
Approaches:
• Exact solution infeasible
• Greedy forward selection
• Backward elimination
• Forward-backward
• Decision tree induction
Backprop is for: Correct Answers Arbitrary feed-forward
topology
Differentiable nonlinear activation functions
Broad class of error function
Backpropagation Correct Answers - Used to calculate
derivatives of error function efficiently
- Errors propagate backwards layer by layer
Backward Elimination Correct Answers 1. Start with complete
SF set (contains all original features)
2. Find feature that, when removed, reduces the filter score least
3. Remove feature from SF set
4. Repeat steps 2-3 until convergence
Basic Decision Tree Correct Answers Decision trees apply a
series of linear decisions, that often depend on only a single
variable at a time. Such trees partition the input space into
cuboid regions, gradually refining the level of detail of a
Specialization Questions With Complete Solutions
1st order Markov models Correct Answers Restricted to
encoding sequential correlation on previous element only
3 categories of machine learning algorithms Correct Answers 1.
Supervised learning which uses labeled data to predict new data
2. Unsupervised learning which finds patterns in unlabeled data
and
3. Reinforcement learning, which trains a model to make
decisions based on rewards and punishments from the
environment
A Bernoulli trial Correct Answers - It is a trial with a binary
outcome, for which the probability that the outcome is 1 equals
p (think of a coin toss of an old warped coin with the probability
of throwing heads being p).
- A Bernoulli experiment is a number of Bernoulli trials
performed after each other. These trials are i.i.d. by definition.
A Latice/Trellis diagram visualizes Correct Answers state
transitions over time
Also good tool to to visualize optimal path through states
(Viterbi Algorithm)
Abstract Essence of ML Correct Answers Representation +
Evaluation + Optimization
accuracy Correct Answers (TP + TN)/ (TP + TN + FP + FN)
,accuracy may not be useful measure in cases where Correct
Answers 1- There is a large class skew
2- There are differential misclassification costs - say, getting a
positive wrong costs more than getting a negative wrong.
3- We are interested in a subset of high confidence predictions
Acquiring emissions Correct Answers Wide range of options to
model ****** probabilities:
- Discrete tables
- Gaussians
- Mixture of Gaussians
- Neural Networks/RVMs etc to mode
All probability theory can be expressed in terms of two rules
Correct Answers - Product rule
- Sum rule
Ancestral sampling Correct Answers is a simple sampling
method well suited to PGNs
ANN feature selection Correct Answers • Artificial Neural
Networks can implicitly perform feature selection
• A multi-layer neural network where the first hidden layer has
fewer units (nodes) than the input layer
• Called 'Auto-associative' networks
Applications of machine learning Correct Answers Self Driving
Cars
Image Classification
Text Classification
Search Engine
,Banking, Healthcare Domain. web searches, photo recognition,
movie recommendations, and spam filtering. It is also being
implemented in industries, including optimizing wind turbines,
healthcare, and factory inspection.
Applications of supervised learning Correct Answers
Classification is a machine learning technique used for tasks like
spam filtering, online advertising, and visual inspection to
classify data into specific categories.
Regression is another machine learning technique used for tasks
like speech recognition, machine translation, and predicting the
position of objects in self-driving cars, where the goal is to
predict numerical values based on input data.
Apriori algorithm Correct Answers Apriori algorithm is a fast
way of finding frequent itemsets
ARFF Correct Answers Attribute-Relation File Format
Artificial Neural Nets Correct Answers • Feed-forward neural
network/Multilayer Perceptron one of many ANNs
• We focus on the Multilayer Perceptron
• Really multiple layers of logistic regression models
Association Rules Correct Answers Reflect items that are
frequently found (purchased) together, i.e. they are frequent
itemsets
• Information that customers who buy beer also buy crisps is e.g.
encoded as:
beer ) crisps[support = 2%, confidence = 75%]
, Attribute subset selection Correct Answers Feature selection
Feature selection is a form of dimensionality reduction in ML,
hence the DM term 'dimensionality reduction' for manifold
projection is problematic.
Approaches:
• Exact solution infeasible
• Greedy forward selection
• Backward elimination
• Forward-backward
• Decision tree induction
Backprop is for: Correct Answers Arbitrary feed-forward
topology
Differentiable nonlinear activation functions
Broad class of error function
Backpropagation Correct Answers - Used to calculate
derivatives of error function efficiently
- Errors propagate backwards layer by layer
Backward Elimination Correct Answers 1. Start with complete
SF set (contains all original features)
2. Find feature that, when removed, reduces the filter score least
3. Remove feature from SF set
4. Repeat steps 2-3 until convergence
Basic Decision Tree Correct Answers Decision trees apply a
series of linear decisions, that often depend on only a single
variable at a time. Such trees partition the input space into
cuboid regions, gradually refining the level of detail of a
