8 Common Machine Learning Algorithms - An Executive's Guide

 

Below is a list of some of the most common machine learning (ML) algorithms. This definition has been taken directly from my MIT class - 

1. Naïve Bayes classifier algorithm is among the most popular learning methods grouped by similarities, which works on the popular Bayes theorem of probability. Naïve Bayes classifier algorithm can be used if you have a moderate or large training dataset, if the instances have several attributes, and if, given the classification parameter, attributes that describe the instances should be conditionally independent. Its applications include sentiment analysis, document categorization, and email spam filtering. 
2. K-means clustering algorithm is a popularly used unsupervised ML algorithm for cluster analysis. The algorithm operates on a given dataset through a predefined number of clusters, k. The output of the k-means algorithm is k clusters with partitioned input data. For instance, let’s consider k-means clustering for Wikipedia search results. The search term “Jaguar” on Wikipedia will return all pages containing the word Jaguar which can refer to Jaguar as a car, Jaguar as a Mac OS version, and Jaguar as an animal. K-means clustering algorithm can be applied to group the web pages that talk about similar concepts. It is used by most search engines, such as Yahoo and Google, to cluster web pages by similarity and identify the “relevance rate” of search results. 
3. Support vector machine (SVM) learning algorithm is a supervised ML algorithm for classification or regression problems where the dataset teaches SVM about classes so that it can classify any new data into different classes by finding a line (hyperplane) which separates the training dataset into classes. The advantages of using SVM include best classification accuracy on the training data, more efficiency for correct classification of the future data, and no overfitting of the data. SVM is commonly used for stock market forecasting by various financial institutions. For instance, it can be used to compare the relative performance of the stocks when compared to the performance of other stocks in the same sector. 
4. Linear regression machine learning algorithm shows the relationship between two variables and how the change in one variable impacts the other. Linear regression is used for estimating real continuous values. The most common examples of linear regression are housing price predictions, sales predictions, weather predictions, employee salary estimations, etc. The basic goal of linear regression is to fit the best line among the predictions. Linear regression is one of the most interpretable machine learning algorithms. It’s fast and requires minimal tuning. 
5. Logistic regression machine learning algorithm is used for classification tasks and not regression problems. Here, “regression” implies that a linear model is fit into the feature space. The odds or probabilities that describe the outcome of a single trial are modeled as a function of explanatory variables. Logistic regression algorithms help estimate the probability of falling into a specific level of the categorical dependent variable based on the given predictor variables. Logistic Regression is a robust algorithm as the independent variables need not have equal variance or normal distribution. Additionally, it does not assume a linear relationship between the dependent and independent variables and, hence, can also handle non-linear effects. The applications of logistic regression include epidemiology to identify the risk factors for diseases and plan accordingly for preventive measures as well as for risk management in credit scoring systems. 
6. Decision tree machine learning algorithm is a graphical representation that makes use of the branching methodology to exemplify all possible outcomes of a decision, based on certain conditions. In a decision tree, the internal node represents a test on the attribute; each branch of the tree represents the outcome of the test, and the leaf node represents a particular class label, i.e., the decision made after computing all the attributes. The classification rules are represented through the path from the root to the leaf node. Decision trees are very instinctual and can be explained to anyone with ease. People from a non-technical background can also decipher the hypothesis drawn from a decision tree, as they are self-explanatory. Decision tree algorithms can handle both categorical and numerical variables and do not require making any assumption on the linearity in the data. Hence, they can be used in circumstances where the parameters are non-linearly related. These algorithms are useful in data exploration and implicitly perform feature selection. In finance, applications of decision trees include banks classifying loan applicants; in medicine, they are used to identify at-risk patients and disease trends. 
7. Artificial neural networks (ANN) algorithms have interconnected non-linear neurons; thus, these machine learning algorithms can exploit non-linearity in a distributed manner. They can adapt free parameters to the changes in the surrounding environment. They learn from their mistakes and make better decisions through backpropagation. ANNs are easy to conceptualize, and they can identify all probable interactions between predictor variables. Financial institutions use ANN machine learning algorithms to enhance their performance in evaluating loan applications and bond rating. Many bomb detectors at U.S. airports use artificial neural networks to analyze airborne trace elements and identify the presence of explosive chemicals. Google uses artificial neural networks for speech recognition, image recognition, and other pattern recognition (handwriting recognition) applications. 
8. K-nearest neighbors (KNN) uses the prediction of continuous values like regression. Distance-based measures are used in k-nearest neighbors to get the closest correct prediction. The final prediction value is chosen on the basis of the k neighbors. The advantages of using k-nearest neighbors include high accuracy; however, better algorithms exist, and ’it’s very useful for non-linear data, as there are no assumptions here.