Logistic Regression is a classical machine learning method. It is a classification algorithm which calculates the probability of a data point belonging to a certain class via a sigmoid function. Due to its simplicity, logistic regression works best on data sets which are correlated as linearly as possible. The data set used in this lab was specially selected due to its linear separability.
Quantum Support Vector Machine (QSVM) is a quantum machine learning method. Similar to Classical SVM, the data is mapped onto a high dimensional feature space where a hyperplane acts as the decision boundary which separates the data into classes. In contrast to SVMs however QSVM employs quantum machine learning principles to facilitate this process. Such an application includes the use of a quantum kernel, where data is encoded into a Hilbert space using a quantum feature map. It is then passed into a classical SVM solver, which performs QP operations to determine the optimal hyperplane.
The "Energy Efficiency" dataset used in this module documents the architectural features of building which contribute to their heating load value. The target variable is the heating load value which is binarized at the median. The four selected features named V1, V3, V4, and V5 correspond to the following architectural characteristics respectively: relative compactness, wall area, roof area, and overall height. This dataset is available on the UC Irvine Machine Learning Repository. It contains 2 targets and 8 features.
"Energy Efficiency" https://archive.ics.uci.edu/dataset/242/energy+efficiency