Fatima Jinnah Women UniversityEmoHealthA Decision Process:With machine learning, you may make predictions or categorizedata. It will create an estimate of a pattern in the data based on some labeled orunlabeled input data.An Error Function:When evaluating a model's prediction, an error function comes inhandy. The correctness of a model can be evaluated by a comparison to knownexamples.An Model Optimization Process:Weights are adjusted if the model's predictions arecloser to the actual data points in the training set than the known example. It isnecessary for the algorithm to repeat the evaluation and optimization process severaltimes in order to acquire a particular degree of accuracy.2.13.4MethodsThe three main types of machine learning classifiers are as follows.Supervised machine learningTo train algorithms that reliably identify data or predict outcomes using labeled datasets is calledsupervised learning, which is also known as supervised machine learning. Model weights areadjusted as input data is fed into it, until a satisfactory match is found for the model. A cross-validation process ensures that the model doesn't fit or underfit the data. Putting spam in aseparate folder from your inbox, for example, aids businesses in dealing with a variety of real-world difficulties on a large scale. supervised learning makes use of neural networks, naivebayes, linear regression, logistic regression, random forests, and support vector machines (SVM).Unsupervised machine learningTo analyze and cluster datasets that haven't been labeled with human assistance, unsupervisedlearning is a technique known as unsupervised machine learning or unsupervised learning. Usingthese algorithms, human researchers are not required to discover previously unknown patterns ordata clusters. Cross-selling tactics, consumer segmentation, and image and pattern recognition allbenefit from its ability to detect similarities and contrasts in data during exploratory analyses.
