چكيده به لاتين
Diagnosis and detection of diseases, especially iron deficiency anemia is being used in various methods, because of the importance of early detection that can help doctors saving patients’ lives. Diagnosis can be facilitated by the machine learning methods and data mining algorithms, which majority of them are supervised and need labeled datasets. The overall goal of the data mining process is to extract information from a dataset and turn it into comprehensible data to help the user making decisions. Discovering knowledge from a large amount of data from patients' records using data mining can lead to improved quality of medical services. These tools can include statistical models, mathematical algorithms, and machine learning methods such as regression and classification. In the process of predicting and diagnosing anemia, mainly classification methods such as k-nearest neighborhood, decision tree, Support vector machine, random forest, and naive Bayes and neural networks have been used. The necessity of data-analysis-based methods is being felt like a vital gap, which is crucial in order to analyze the collected data and elicitate the useful knowledge out of it. Therefore, recognition of eligible features and factors for iron deficiency anemia can be the advantage of mentioned model. Thus, in this dissertation, blood data has been taken from a laboratory in the 14th district of Tehran and after preprocessing and normalizing the data, algorithms were used to predict the disease. If individuals are healthy, the diagnosis is done with the utilization of CBC tests and does not need an appointment with a doctor.
