Background: In rural areas of Indonesia, where parasitological tests are not always available, a knowledge of significant symptoms and non-symptom-related factors is crucial for clinical diagnosis of malaria in order to take appropriate actions for treatment. Malaria diagnosis at early stages can avoid a progression in severity of the conditions and possible transmission of the disease. With insufficient knowledge of significant malaria symptoms and non-symptom-related factors for clinical diagnosis of malaria, medical doctors may presumptively treat febrile patients, which leads to an increase in resistance to antimalarial drugs.Methods: The study involves five stages. First, medical records are collected from patients suffering from malaria and other febrile diseases such as dengue fever, typhoid fever, respiratory tract infection, dyspepsia, pneumonia, common cold and gastritis. The records are obtained from two public hospitals in endemic regions in Indonesia’s East Nusa Tenggara Province, East Sumba and Lembata. Second, interviews with eight Indonesian medical doctors are conducted to understand the symptoms and non-symptom-related factors they use for clinical diagnosis of malaria. Third, symptoms and non-symptom-related factors from 472 collected medical records are selected and ranked using the correlation matrix method. Fourth, the 17 feature sets extracted from the ranked features are used to develop and compare various malaria classifiers using three machine-learning techniques, including logistic regression, support vector machines and k-nearest neighbors. Fifth, the most significant symptoms and non-symptom-related factors are identified for malaria diagnosis using the malaria classifiers developed. In addition, further experiments and comparisons are conducted to examine the performance of malaria classifiers using various feature sets.Results: (a) Out of 17 features identified from medical records, six significant symptoms (duration of fever, headache, nausea and vomiting, heartburn, pain and severe symptoms), and two non-symptom-related factors (history of malaria and age) contribute most to malaria diagnosis. These significant features can be used for malaria diagnosis with an average 85% accuracy. (b) As a symptom, duration of fever is more significant than temperature or fever for distinguishing malaria from other febrile diseases. (c) Trios malaria symptoms (fever, shivering and sweating) are shown to be less significant than other symptoms in endemic regions.Conclusions: The findings of this study can be used to raise the public awareness of significant malaria symptoms and non-symptom-related factors within societies vulnerable to malaria. These findings can also be used for helping unexperienced medical doctors with clinical diagnosis of malaria especially in rural areas, where the availability of parasitological tests is scarce.