Machine Learning Part 1

Machine learning uses algorithms to find patterns in data, and then uses a model that
recognizes those patterns to make predictions on new data.

Machine learning may be broken down into -

1. Supervised learning algorithms use labeled data - Classification, Regression.
2. Unsupervised learning algorithms find patterns in unlabeled data - Clustering, Collaborative Filtering, Frequent Pattern Mining
3. Semi-supervised learning uses a mixture of labeled and unlabeled data.
4. Reinforcement learning trains algorithms to maximize rewards based on feedback.

Classification - Mailing Servers like Gmail uses ML to classify if an email is Spam or not based on the data of an email: the sender, recipients, subject, and message body. Classification takes a set of data with known labels and learns how to label new records based on that information. For example- An items is important or not. A transaction is fraud or not based upon known labeled examples of transactions which were classified fraud or not.

Regression - Like Linear Regression predicts a numerical value. Example - A person car got damaged and he claimed 1000$ to insurance company. And, historically company has seen or settled insurance amounts up to 800$. Thus, amount of 1000$ can be a fraud amount.

Clustering - Google News uses a technique called clustering to group news articles into different categories based on title and content. Clustering algorithms discover groupings that occur in collections of data.

Collaborative Filtering - Amazon uses a machine learning technique called collaborative filtering (commonly referred to as recommendation) to determine which products users will like, based on their history and similarity to other users. For example - Recommendation : A person buying this Shampoo also purchased Comb & Hair Conditioner.

Supervised algorithms use labeled data - in which both the input and target outcome, or label, are provided to the algorithm.

Supervised Learning is also called predictive modeling or predictive analytics, because you build a model that is capable of making predictions.

Classification and Regression take a set of data with known labels and pre-determined features and learns how to label new records based on that information. Features are the “if questions” that you ask. The label is the answer to those questions. For example - if it is a Fruit, Red or Green in color, Sweet in Taste, Farmed in Cold Regions, Having seeds but not more than 5... then it must be an Apple (Label).
So, if a fruit is orange in color it might not classify as an Apple. It might be orange.

Multiple linear regression models the relationship between two or more “Features” and a response “Label.” For example, if we wanted to model the relationship between the amount of fraud and the age of the claimant, the claimed amount, and the severity of the accident, the multiple linear regression function would look like this:

Amount Fraud = intercept + (coefficient1 * age) + (coefficient2 * claimed Amount) +
(coefficient3 * severity) + error.

Logistic regression measures the relationship between the Y “Label” and the X “Features” by estimating probabilities using a logistic function. The model predicts a probability, which is used to predict the label class. For example - Label can be "Probability of Fraud" and Features can be "Transaction amount, merchant type, time location, and time difference since last transaction"

Classification:Logistic regression, Decision tree classifier, Random forest classifier, Gradient-boosted tree classifier, Multilayer perception classifier, Linear Support Vector Machine, Naive Bayes

Regression:Linear regression, Generalized linear regression, Decision tree regression, Random forest regression, Gradient-boosted tree regression, Survival regression, Isotonic regression.

--------------

Unsupervised learning, also sometimes called descriptive analytics, does not have labeled data provided in advance. These algorithms discover similarities, or regularities, in the input data. An example of unsupervised learning is grouping similar customers, based on purchase data.

Clustering- an algorithm classifies inputs into categories by analyzing similarities between input examples. Ex- Grouping similar customers, Text categorization, Network Security Anomaly detection (anomalies find what is not similar, which means the outliers from clusters)

The k-means algorithm groups observations into k clusters in which each observation belongs to the cluster with the nearest mean from its cluster center.

Clustering with unsupervised learning is often combined with supervised learning in order to get more valuable results.

Collaborative Filtering - Frequent Pattern Mining, Association, Co-Occurrence, Market Basket Recommendations - Frequent pattern or association rule mining finds frequent co-occurring associations among a collection of items, such as products often purchased together.  Ex - beer and diaper story.

Clustering: k-means, Latent Dirichlet allocation (LDA), Gaussian mixture model (GMM)
Collaborative Filtering: Alternating least squares (ALS)
Frequent Pattern Mining: FP-Growth Algorithm

-----------
Deep learning is the name for multilayered neural networks, which are networks composed of several “hidden layers” of nodes between the input and output. Each node takes input data and a weight and outputs a confidence score to the nodes in the next layer, until the output layer is reached, where the error of the score is calculated. With backpropagation inside of a process called gradient descent, the errors are sent back through the network again and the weights are adjusted, improving the model. This process is repeated thousands of times, adjusting a model’s weights in response to the error it produces, until the error can’t be reduced any more.

During this process the layers learn the optimal features for the model, which has the advantage that features do not need to be predetermined. However, this has the disadvantage that the model’s decisions are not explainable.

Deep learning libraries or frameworks that can be leveraged with Spark include: BigDL, Spark Deep Learning, Pipelines, TensorFlowOnSpark, dist-keras, H2O Sparkling Water, PyTorch, MXNet, Caffe

Part2 - https://techdevins.blogspot.com/2019/12/machine-learning-part-2.html