Background and purpose: High rate of mothers and infants’ death and preterm birth are amongst major public health problems worldwide. The aim of this study was to identify, evaluate and rank the factors responsible for low birth weight using data mining techniques and also investigating the impact of predictor variables and developing a decision support system which could help physicians to make better treatment decisions at the birth of low weight infants.

Materials and methods: Relevant information was collected from Imam Ali Hospital affiliated with Zahedan University of Medical Sciences in 2013 including smoking, the age of mothers, etc. Different data mining algorithms were applied for modeling. Data analysis was performed in Clementine software.

Results: The variables that were very influential in predicting the low weight of infants at birth were mother’s weight (100%), mother’s age (98%), the number of doctor visits in the first trimester of pregnancy (45.86%), and previous preterm delivery (43.11%). Other variables poorly influenced the prediction.

Conclusion: The findings revealed some relationships between the low weight of infants at birth and mother’s weight, mother’s age, number of doctor visits in the first trimester of pregnancy, previous preterm delivery, high blood pressure, race, uterine irritability, and smoking. The accuracy of prediction improved via data mining techniques compared to logistic regression. Classification tree could determine the low weight of infants at birth well and random forest technique had an important role in making the diagnosis.