The goal of this study is to summarize the seven most frequent types of categorization algorithms.

Data can be classified in two ways: structured and unstructured. Classification is a technique for categorizing data into a set of categories. The main goal of a classification problem is to determine which category or class to which new data will belong.

Logistic Regression

Logistic regression is a classification machine learning algorithm. A logistic function is used to model the probability of the probable outcomes of a single trial in this technique. Logistic regression was created for this goal (classification), and it's especially good for figuring out how numerous independent factors affect a single outcome variable. The downside is that it only works with binary variables, requires that all predictors are independent of one another, and assumes that the data is free of missing values.

Naïve Bayes

The Naive Bayes algorithm is based on Bayes' theorem and assumes that every pair of features is independent. Many real-world situations, such as document classification and spam filtering, benefit from Naive Bayes classifiers.

To estimate the required parameters, this technique requires a limited amount of training data. When compared to more advanced algorithms, Naive Bayes classifiers are extremely fast.

Stochastic Gradient Descent

Stochastic gradient descent is a straightforward and highly efficient method for fitting linear models. It's particularly beneficial when there are a lot of samples. For classification, it provides a variety of loss functions and penalties. The disadvantage is that it necessitates the use of several hyper-parameters and is sensitive to feature scaling.

K-Nearest Neighbors

Neighbor-based classification is a sort of lazy learning because it doesn't try to build a general internal model and instead just saves instances of the training data. The classification is determined by a simple majority vote of each point's k nearest neighbors. This technique is straightforward to use, robust to noisy training data, and successful when dealing with massive amounts of data. It is necessary to estimate the value of K, and the calculation cost is considerable because each instance must be distanced from all of the training samples.

Decision Tree

A decision tree generates a set of rules that can be used to categorize data given a set of attributes and their classes. The Decision Tree is easy to comprehend and visualize, requires minimal data preparation, and can handle both numerical and categorical data(Dr Iain Brown 2015). Decision trees can produce complicated trees that are difficult to generalize, and they can be unstable since slight changes in the data might result in the generation of an entirely different tree.

Random Forest

The random forest classifier is a meta-estimator that fits several decision trees on different sub-samples of datasets and utilizes average to improve the model's predictive accuracy and control over-fitting. The size of the sub-sample is always the same as the size of the original input sample, but the samples are generated with replacement. In most circumstances, a reduction in over-fitting and a random forest classifier are more accurate than decision trees. Real-time prediction is slow, and the algorithm is complicated to implement.

Support Vector Machine

The training data is represented as points in space split into categories by a clear gap as broad as possible in a support vector machine. New examples are then mapped into the same space and classified according to which side of the gap they fall on.

It works well in high-dimensional areas and only utilizes a small number of training points in the decision function, making it memory-friendly.

The algorithm does not offer probability estimates directly; instead, they are derived through a time-consuming five-fold cross-validation procedure.