Explore chapters and articles related to this topic
Discrimination of Healthy Skin, Superficial Epidermal Burns, and Full-Thickness Burns from 2D-Colored Images Using Machine Learning
Published in Qurban A. Memon, Shakeel Ahmed Khoja, Data Science, 2019
Aliyu Abubakar, Hassan Ugail, Ali Maina Bukar, Kirsty M. Smith
ML is a subset of artificial intelligence (AI) that provides how to make machines intelligent and enable them to act like humans [27–29]. The terms AI and ML are sometimes used interchangeably; however, the terms differ but with a strong relationship. AI is a concept of making machines intelligent, while ML is a technique of how to achieve AI. ML basically can be grouped into two main categories: supervised ML and unsupervised ML. In supervised learning technique, machines are trained with labeled data so as to learn to map an input data with the corresponding output. The labeled data means a known data (i.e., a problem at hand with the known solution), and the idea is to enable machines to learn the relationship between data and output by learning unique representations that associated the input with the output. The goal is to enable accurate prediction of unseen data when presented to the machine without human intervention. The process of learning that involves only input data with no output information is called unsupervised learning. In this type of learning, machines are allowed to figure out and group data based on the similarity of representations. Unsupervised learning is considered more of true AI by some researchers than supervised learning approach, because during the learning process, there is complete absence of human intervention to guide the learning process. Some examples of supervised ML algorithms are support vector machines (SVMs) and artificial neural networks, while clustering is an example of unsupervised learning strategy.
Machine Learning Basics
Published in Fei Hu, Qi Hao, Intelligent Sensor Networks, 2012
Krasimira Kapitanova, Sang H. Son
Collecting labeled data is resource and time consuming, and accurate labeling is often hard to achieve. For example, obtaining sufficient training data for activity recognition in a home might require 3 or 4 weeks of collecting and labeling data. Further, labeling is difficult not only for remote areas that are not easily accessible, but also for home and commercial building deployments. For any of those deployments, someone has to perform the data labeling. In a home deployment, the labeling can be done by the residents themselves, in which case they have to keep a log of what they are doing and at what time. Previous experience has shown that these logs are often incomplete and inaccurate. An alternative solution is to install cameras throughout the house and monitor the activities of the residents. However, this approach is considered to be privacy-invasive and therefore not suitable.
Application of Artificial Intelligence in Image Processing
Published in Nedunchezhian Raju, M. Rajalakshmi, Dinesh Goyal, S. Balamurugan, Ahmed A. Elngar, Bright Keswani, Empowering Artificial Intelligence Through Machine Learning, 2022
Fei-Fei Li, an eminent professor and head of the AI lab at Stanford University, launched ImageNet in the year 2009.6 ImageNet is basically a database containing labeled images. By the end of 2017, the ImageNet consists of more than 14 million labeled images which were made available to students, educators, and researchers. Labeled data are in general those that are to be trained using the supervised learning process. Images were labeled and organized using the WordNet, which is a lexical database of English words where verbs, adverbs, nouns, and adjectives are sorted and grouped by synonyms called synsets.
Machine Learning Techniques in Adaptive and Personalized Systems for Health and Wellness
Published in International Journal of Human–Computer Interaction, 2023
Oladapo Oyebode, Jonathon Fowles, Darren Steeves, Rita Orji
In supervised learning, labelled data is used to train a model which then predicts the label (or class) for new data (Liu, 2011). In other words, the training data contain both the inputs and the expected outputs; hence the model can learn to predict the actual outputs of new (unseen) inputs. The vast majority of the reviewed papers (n = 82, 94%) reported the use of supervised learning technique for classification and regression tasks. For classification tasks, models are used to predict a binary class (such as occurrence or absence of a fall (Mrozek et al., 2020)) or multiclass (e.g., at-risk health conditions (Asthana et al., 2017)); however, for regression tasks, models predict a continuous value (e.g., blood pressure (P. Chiang & Dey, 2019)). To improve readability, we further categorized the supervised learning methods into classical ML, ensemble learning, and deep learning.
A Systematic Review of Recent Machine Learning Techniques for Plant Disease Identification and Classification
Published in IETE Technical Review, 2023
Considerable studies and surveys have been performed in the past few years to compare the analysis of classification methods in machine learning. The branch of machine learning is broadly classified into supervised and unsupervised learning [23]. Supervised learning deals with the labeled data, where the machine is trained by providing input and their corresponding correct output, and it uses this “labeled” data to predict the output of unknown input. Classification and Regression Problems come under Supervised learning. In contrast, Unsupervised Learning learns from experience. The idea behind unsupervised learning is to map a pattern or structure from the unlabeled data without any supervision. Further Association and Clustering are the categories of unsupervised learning. The comparative study shows that the Convolution Neural network (CNN) [34,35], K-Nearest Neighbor(K-NN) [36,37], Artificial Neural Network (ANN) [33,38], and Support Vector Machine (SVM) [39,40] are the ones that are used at most and provide accurate prediction for the Classification of plant disease type. Further Table 3 provides a comparative analysis that explains four majorly used Machine learning techniques to classify crop diseases. It provides a brief review of each machine learning classification model used for researching different crops and the details of the data set on which the research study is performed, and their corresponding classification results based on various performance measures.
A perspective on machine learning in turbulent flows
Published in Journal of Turbulence, 2020
Sandeep Pandey, Jörg Schumacher, Katepalli R. Sreenivasan
Machine learning can be classified into three big categories – unsupervised, semi-supervised and supervised machine learning [5,18]. Unsupervised machine learning extracts features in (high-dimensional) data sets without pre-labelled training data. These techniques are established already and known for decades. They comprise, for example, clustering such as the k-means or spectral clustering and dimensionality-reduction techniques, such as the well-known Proper Orthogonal Decomposition (POD) [19] or Dynamic Mode Decomposition (DMD) [20]. The new element is the enormous amount of data that can be handled now. In contrast, supervised machine learning requires labelled data for the training of the ML algorithm. Needless to say, semi-supervised learning combines aspects of supervised and unsupervised learning methods.