Hypertension is a major medical problem that may lead to cardiovascular diseases. It is usually assessed through a periodic but systematic measurement of systolic and diastolic blood pressure. Since the electrocardiogram (ECG) is one of the most widely used diagnostic tools, it could be used for the initial evaluation of a patient suspected to have hypertension, if this information is actually contained in the ECG. We show that through the use of machine learning that this is indeed the case, i.e., that use of ECG may be used in detecting hypertension in a population without cardiovascular disease. The machine learning methods used involve logistic regression, k-nearest neighbors, random forest, and gradient boosting. Through this analysis, we find the basic clinical and ECG features that combined may give information on the hypertensive state of individuals and thus assist in early diagnosis and treatment.