Search

Background: Response variables in most medical and health-related research have an ordinal nature. Conventional modeling methods assume predictor variables to be independent, and consider a large number of samples (n) compared to the number of covariates (p). Therefore, it is not possible to use conventional models for high dimensional genetic data in which p > n. The present study compared the predictive performance of decision trees, ordinal forest, and L1 penalized continuation ratio regression. Materials and Methods: In the present study, three data sets were used. The B-cell data contained 12,625 gene expression data related to 128 patients with four ordinal levels of response variables. The HCC data related to liver cancer included 1469 genes of 56 patients with three ordinal levels of response variables. The Heart data contained information of five variables in 294 patients undergoing angiography with five ordinal levels of response variables. The performance of the methods was compared based on the same training and test datasets using indicators such as accuracy, gamma, and kappa. Results: For two high-dimensional data sets, the ordinal forest model had a higher predictive ability while for the low-dimensional data set, the L1 penalized continuation ratio model had a better predictive performance . Conclusion: The selection of the best prediction model depends on the data set, and for each data, different methods should be considered to achieve the best model. [ABSTRACT FROM AUTHOR]

In decision making, timeliness is one of the useful information features that have increasingly attracted the attention of regulatory agencies. This paper examines factors which are related to the timeliness of annual reporting financial statements in Tehran stock exchange listed firms. The effect on financial reporting timeliness of factors such as the ratio of the investment institutes, major shareholders and institutional shareholders ownership, the ratio of the shares held by the CEO and board, the quantity of the majority shareholders, and the ratio of the non-executive members, duality, big auditor, auditor change, size, age, ROE, good news, consolidate financial reporting and the quality of the costing system, opinion of the independent auditor and industry during the years 2008 to 2011 have been studied. Neural networks and decision trees techniques is utilized for data analysis. The Results show that the institutional shareholders ownership, quantity of the majority shareholders, industry, big auditor, opinion of the independent auditor, ratio of the non-executive members, consolidate financial reporting and ROE have the greatest effect on quality of timeliness, respectively. Also average accuracy of neural networks technique is higher than decision trees technique.

Decision trees as one of the data mining techniques, is used in credit scoring of bank customers. The main problem is the construction of decision trees in that they can classify customers optimally. This paper proposes an appropriate model based on genetic algorithm for credit scoring of banks customers in order to offer credit facilities to each class. Genetic algorithm can help in credit scoring of customers by choosing appropriate features and building optimum decision trees. Development process in pattern recognition and CRISP process are used in credit scoring of customers in construction of this model. The proposed classification model is based on clustering, feature selection, decision trees and genetic algorithm techniques. This model select and combine the best decision tree based on the optimality criteria and constructs the final decision tree for credit scoring of customers. Results show that the accuracy of proposed classification model is more than almost the entire decision tree models compared in this paper. Also the number of leaves and the size of decision tree i.e. its complexity is less than the other models.

In the present study, the relationship between the distribution of Astragalus verus, climate, soil and topography in Isfahan province was investigated using the CART (Classification and Regression Trees) non-parametric regression. According to the vegetation types dominated by Astragalus verus, 287 sites were selected using stratified-random sampling and 106 presence points were recorded. Non-normally distributed soil variables were normalized according to the skewness type by data transformation. For variables that were not normalized by the data transformation, the inverse distance weighted and for normally distributed variables, kriging methods were used for mapping. To investigate the spatial continuity of these variables, the best variogram was selected. Using the principle component analysis (PCA) and correlation matrix, the most effective factors on distribution were clay percentage, mean temperature of the coldest (Bio11), and driest (Bio9) quarter, minimum temperature of the coldest month (Bio6), annual mean temperature (Bio1), saturated moisture and organic carbon, respectively. Model evaluation using replacement method and using independent data indicated high accuracy of the model. According to the optimum threshold of 0.41 and fuzzy-based habitat suitability map, the suitable habitat for the species was 34.5 % of the area. The results are used in sustainable rangeland management, conservation and their restoration. [ABSTRACT FROM AUTHOR]

Background and Aim: Diabetes may increase the incidence of tooth decay due to dry mouth and high blood sugar levels. Identifying the factors influencing oral health behaviours in diabetic patients is thus an essential step toward promoting oral and dental health. As a result, this study aimed to predict oral health in people with diabetes and compare them to healthy people. Material and Methods: The available sampling method was used to conduct this study from 2021 to 2022. The study group consisted of 261 persons (men and women), 131 of whom were healthy and 130 of whom were unhealthy (diabetic), and information was gathered through a questionnaire, medical records, and an examination. These people looked at six variables: age, gender, decayed teeth, extracted teeth, filled teeth, and oral health index. Using the Spss Modeler program, two decision tree methods and a support vector machine and spss Modeler soft ware were used. Results: The most important findings of decision tree analysis are 1- If the person's age is less than or equal to 37 years, then the person is 100% healthy. 2- If the age is over 37 years and the number of decayed teeth is less than the average of 7, and we do not have any extracted teeth, there is an 82% chance of diabetes. If the age is over 37 and the number of decayed teeth is less than the average of 7, and the number of extracted teeth is more than 1, then people under the age of 49 with an OHI index greater than 0.9 are 100% diabetic. Also, the total accuracy of the linear support vector machine is 70.73%, which indicates that decayed teeth with the least amount of weight have little effect on diabetes or health. Conclusion: Decision tree algorithms and support vector machines could predict oral and dental health in diabetic patients. [ABSTRACT FROM AUTHOR]

Today, companies and organizations are aware of the fact that customer retention leads to greater profitability. Increasing competition causes the rate of customer churn to grow. Therefore, studying the features influencing the tendency of customer churn is important. In the present study, a hybrid model based on the data mining approach is presented to analyze the features of churn customers. In the first step, the feature selection node has been used to identify the features with higher importance and remove redundant items. Then, the C5.0 Decision Tree and Bayesian network were used to classify the customers into two groups, turning customers, and non-turning customers. These are data mining techniques and terms that can help in forecasting. Finally, the proposed model has been implemented in the chain store industry as a case study. Key findings indicate that both the decision tree model and the Bayesian network can predict churn customers with different accuracies, the area under the receiver operating curve in the decision tree model is greater than the Bayesian network model and it has better performance. The results indicate the optimal efficiency of the proposed method. In addition, the results show that three features of gender, marital status, and age from the set of demographic characteristics and five factors of average monthly income level, number of purchases per month, the share of online shopping, how to get to know the store, type of market-related to customer transaction records are among the most effective factors. [ABSTRACT FROM AUTHOR]

The stability of curcumin and its encapsulation efficiency in emulsion are among the most important factors determining its bioavailability and absorption in the body. For this purpose, parameters affecting these two factors, including time, pH, and cellulose concentration, were used as input variables in the present study. Curcumin stability and encapsulation efficiency were used as response variables in artificial neural networks and decision tree algorithms. In this regard, cellulose nanocrystal obtained from acid hydrolysis was used to prepare curcumin Pickering emulsion with oil:water ratio of 5:95 and cellulose concentration 1, 1.5, 2, 2.5, and 3% (w/v) and the encapsulation efficiency and stability of curcumin were measured during 8 days. The results showed that the encapsulation efficiency significantly increased as cellulose concentration increased. Encapsulation efficiency at pH 7 was higher than at pH 2 (p≤0.05). The curcumin content in the emulsions prepared at pH 2 drastically decreased during storage, while it was less degraded in emulsions at pH 7 after 8 days of storage. The modelling results for curcumin stability and encapsulation efficiency based on R2and RMSE% statistics showed that MLP 3 - 6 -1 (R2=0.99; RMSE%=10.00) and RBF 2 - 6 -1 (R2=0.99; RMSE %=9.99) were had more accuracy than other models. Finally, the results showed that the artificial neural network algorithm performed better than the decision tree in predicting cellulose Pickering emulsions' encapsulation efficiency and curcumin stability. [ABSTRACT FROM AUTHOR]

Introduction: Methods of detecting Coronary Artery Disease (CAD) are often prone to error and are also expensive and painful for the patient; therefore, the development and introduction of accurate machine learningbased methods for diagnosing this condition is of high importance. This research aimed to help detect coronary artery disease using the Harris Hawks Optimization (HHO) algorithm and machine learning techniques. Methods: In this research, a novel approach based on feature selection was employed through a combination of HHO and machine learning techniques such as a Decision Tree (DT) and k-Nearest Neighbors algorithm (k-NN). To evaluate the proposed approach, we used two datasets (Cleveland & ZAlizadeh-Sani) with medical records of 303 patients, and the evaluation was conducted by means of python 2016. Results: On the basis of the findings of this research, feature selection by using the Harris hawks optimization algorithm in combination with machine learning methods resulted in an increase in the accuracy of the results in such a way that in the case of Z-Alizadeh-Sani dataset, the percentage of accuracy in combination with a decision tree was equal to 0.98 and in combination with the k-nearest neighbors algorithm was equal to 0.78. Furthermore, the results of the Cleveland dataset showed that using the HHO in combination with a decision tree led to 88 percent accuracy and in combination with the k-nearest neighbors algorithm led to 77 percent accuracy. However, in the case of using all of the features (HHO only mode), accuracy was lower in all cases. Therefore, the HHO algorithm in combination with the decision tree was able to achieve the highest accuracy in diagnosing CAD in the feature selection mode compared to using all of the features. Conclusion: The results from this study showed that the Harris hawk optimization algorithm in combination with machine learning techniques can have a positive role in the process of selecting effective features in diagnosing coronary artery disease. [ABSTRACT FROM AUTHOR]

The objective of this study is to identify and ranking of factors affecting detecting the financial statement frauds using the judgment technique based on the Analytic Hierarchy Process and data mining techniques techniques. The population of the study comprised of senior auditors, supervisors, senior supervisors, audit manager and partner of the audit institute employed in audit institutes member and also companies listed in the Tehran Stock Exchange. In order to the research goal, 56 questionnaires and 109 Listed for the year 2012-2017 and analyzed. Based on the technique of judgment, the pressure dimension of the first priority, opportunity, second factor and rationalization are ranked as the third effective factor on the detection of fraud. These results are different with other techniques. Empirically, the ANNs and CART approaches work with the training and testing samples in a correct classification rate of 98/65% (ANNs) & 91.5% (CART) and 69/79% (ANNs) & 69.10% (CART), respectively, which is more accurate than the logistic model that only reaches 72.32% and 88.10% of the correct classification in assessing the fraud presence. In addition, type II error of CART drops significantly to 58.18% from 72.7% and 55.6% compared to the ones using ANNs and logistic models. According to the accuracy index, the decision tree model is more efficient than other models; therefore, among the data mining techniques, the weight of each of the input variables of the decision tree is the basis for the final ranking of the research variables

Identification and investigation of changes in the area under cultivation of various crops seem to be essential for the management supply of crop production. In this study, r to identify and investigate change of the area under cultivation in major crop Hoseynabade Mishmast region in Qom province, we used the time series images of OLI and ETM sensors of landsat 8 and 7satellites, according to the crop calendar of this region. By using the vegetation index (NDVI) in the decision tree algorithm, the thresholds of this index were adjusted according to the major crops of this region; then a map of the cultivation pattern of the crop of this region was prepared. In order to evaluate the results, the statistics of the provinces agricultural jihad were used during 2005, 2009, 2014 and 2019 crop years. The results showed that by using the threshold of NDVI index, crops in this region in 2005 included wheat and barley and alfalfa, and their areas had an error of 17/1 and 6/1 percent in comparison with the statistics of agricultural Jihad, respectively; in 2009, wheat and barley, alfalfa and corn had an error of 0/5, 9/6 and 0/1 percent. Also, in 2014, wheat and barley, alfalfa, corn and sophie crops had an error equal to 4/9, 0.4, 11/4 and 2/4 percent, and the same crops in 2019 had an error 0/04, 11/6, 1/4 and 17/5 percent; that error was not significant. According to the results, the appropriate efficiency NDVI index in estimating crop cultivation area was determined by their phenology. Also, in 2009 and 2014, corn and sophie crops were added to the regions crops, and the area under crops cultivation in 2019 was increased, as compared to 2014. [ABSTRACT FROM AUTHOR]

Background and Aim: Today, we are witnessing tremendous advances in medical data mining. The data, by analyzing and discovering the relationships between them, can lead to algorithms that help us prevent or treat many diseases. Meanwhile, genetic diseases have attracted a large part of the attention of the medical world because the birth of children with genetic disorders imposes a great financial, psychological and emotional burden on society. Therefore, the aim of this study was to present an algorithm as a secondary screening test before performing cell and molecular tests. Material and Methods: In this study, we studied 1000 cases of pregnant women who were in moderate or high-risk group after screening tests. Their clinical data were stored, missing data was deleted, and records were integrated. Then, using Clementine software, data mining and data correlation were performed, and finally a suitable algorithm was performed for diagnosing the disease. Genetic mutations were identified as well. Results: By applying five algorithms of neural networks, support vector machine, binary decision tree, multiple decision tree and logistic regression on the data, it was found that the neural network algorithm with 97.522% accuracy had the highest success rate in diagnosis of genetic-chromosomal diseases before birth. Conclusion: The use of genetic algorithm as a screening test causes less people to be candidates for costly and dangerous cellular and molecular tests and can be used as a tool to help detect the disease. [ABSTRACT FROM AUTHOR]

Background and Objective: Paying attention to the health of workers as a significant part of the population is important as they play an important role in the development of the society, which also has caught the attention of government officials and World Health Organization (WHO). Based on the rules and regulations of workers in different occupations, each year they must undergo certain medical tests and examinations to ensure they have sufficient health to perform their duties. This study aimed to predict the results of examinations, extraction of knowledge and identifying patterns and agents that affect workers' health. Materials and Methods:This was a descriptive-analytic study conducted in Tehran among 1267 employees of various occupations who participated in annual examinations of labor medicine in 2017 and 80 variables related to their health and occupational and family background were collected during the examinations. Due to the size and type of data, the C5.0 decision tree method was used to perform data mining and discovery process. Results: Using the C5.0 decision tree, a model with accuracy of 99.05% was introduced. According to this model, variables with the greatest impact on the health of the employees were identified. Hearing status, especially hearing loss at frequencies of 6000 and 4000 Hz, had the greatest impact on the results of employee health examinations. Conclusion: According to the extracted patterns and identification of determinants that had the greatest impact on the result of medical examinations, it is possible to control the specified factors to improve the health of workers. [ABSTRACT FROM AUTHOR]

Background: Data mining is known as a process of discovering and analysing large amounts of data in order to find meaningful rules and trends. In healthcare, data mining offers numerous opportunities to study the unknown patterns in a data set. These patterns can be used to diagnosis, prognosis and treatment of patients by physicians. The main objective of this study was to predict the level of serum ferritin in women with anemia and to specify the basic predictive factors of iron deficiency anemia using data mining techniques. Methods: In this research 690 patients and 22 variables have been studied in women population with anemia. These data include 11 laboratories and 11 clinical variables of patients related to the patients who have referred to the laboratory of Imam Hossein and Shohada-E- Haft Tir hospitals from April 2013 to April 2014. Decision tree technique has been used to build the model. Results: The accuracy of the decision tree with all the variables is 75%. Different combinations of variables were examined in order to determine the best model to predict. Regarding the optimum obtained model of the decision tree, the RBC, MCH, MCHC, gastrointestinal cancer and gastrointestinal ulcer were identified as the most important predictive factors. The results indicate if the values of MCV, MCHC and MCH variables are normal and the value of RBC variable is lower than normal limitation, it is diagnosed that the patient is likely 90% iron deficiency anemia. Conclusion: Regarding the simplicity and the low cost of the complete blood count examination, the model of decision tree was taken into consideration to diagnose iron deficiency anemia in patients. Also the impact of new factors such as gastrointestinal hemorrhoids, gastrointestinal surgeries, different gastrointestinal diseases and gastrointestinal ulcers are considered in this paper while the previous studies have been limited only to assess laboratory variables. The rules of the decision tree model can improve the process of diagnosing and treatment of the patients with iron deficiency anemia and reduce their costs.

Background and Objectives: According to wide mass data collection at medical centers and proper use of it in order to diagnosis of a malady needs to relevant tools and medical science for data analyzing. Infertility diagnosis studied by data mining techniques. Methods: All information had been extract from patientchr('39')s documents of ACECR Center for Infertility Treatment, Qom Branch; 700 sample were selected among 14,242 cases in 15 years of age, duration of infertility, family connections, infertility, family, job, male, female menstrual cycle type, hirsutism, galactorrhea, amenorrhea, cause of infertility, female body mass index, smoking and semen tests were used. The prediction algorithms C5.0, C & R tree, CHAID and K-means clustering algorithm to determine the optimal number of clusters Davis - Buldian used. Results: According to the accepted model, the error is less CHAID algorithm, the most important factor in infertility in the female body mass index, age, disease, hirsutism, infertility, family, illness, galactorrhea, the amount of sperm per milliliter, duration of infertility, old man, were consanguineous couples. According to this model, most of the menchr('39')s wear agents were identified. Conclusion: In this study, the effect of female infertility factors predicted.

Objective This article aims to forecast Iran Cooperative Development Bank's profit/loss using a two-stage collective learning method. Employing machine learning for profit and loss prediction is a novel approach to numerical computations, aligning with the article's goal of leveraging big data. Hence, we employ a two-stage collective learning method utilizing Support Vector Machines, Decision Trees, and Weighted Averaging models for learning, testing, and predicting the profit/loss of the Cooperative Development Bank. Methods In the initial stage of machine learning, Support Vector Machines, and Decision Trees serve as the base models, while the second stage employs a weighted averaging approach. The combination of base learning models was utilized to initially test the prediction process on the performance of 35 branches of the Cooperative Development Bank nationwide from 2012 to 2021. The two-stage machine learning method relies on the utilization of 12 variables grouped into 5 factors for the task. These variables encompass interest-free loans and short-term deposits, deposit interest expenses, total customer deposits, income from facilities and investments, common income, cash balances, administrative expenses, long-term deposits, non-common income, asset depreciation expenses, and the bank's size. Relevant data for the study was obtained from the financial statements of 35 branches of the Cooperative Development Bank spanning the years 2012 to 2021. To minimize prediction errors and facilitate ratio comparisons, all indicators mentioned were adjusted relative to the total value of the bank's assets. Furthermore, to minimize prediction errors and enhance the comparability of ratios, all the mentioned indicators were adjusted in proportion to the total value of the bank's assets. In the initial stage of applying this method, two machine learning models - Support Vector Machines and Decision Trees - were employed, followed by a weighted averaging approach in the second stage. Results This article contrasts linear regression with machine learning approaches for predicting the Cooperative Development Bank's actual profit/loss. The results reveal notably highperformance accuracy, evidenced by an MAE metric of 5.66 and an MSE metric of 620.34. Additionally, the correlation between training data and predictions from the twostage collective machine learning stands at 0.9977. Following this method's performance assessment, it is subsequently employed to predict the Cooperative Development Bank's profit/loss for the years 2022 to 2027. Conclusion The results, showcasing the high efficiency of the two-stage collective machine learning method, suggest that managers can employ these approaches for profit/loss prediction in banks. Based on this method, and under normal conditions without abnormal or nonnormal circumstances, the obtained results indicate a prospective decrease in the accumulated losses of the Cooperative Development Bank in future years, leading to an ultimate increase in profits by the year 1405. The results of data analysis reveal that the average ratio of the net profit or loss of the bank to its total assets has been computed in a manner that reflects the average profitability of the bank branches over the research period. [ABSTRACT FROM AUTHOR]

Background: Colorectal cancer is one of the most common cancers in terms of morbidity and mortality worldwide. A lot of research have been done in this field in Iran and worldwide, which have positive results. The aims of this study were firstly doing a statistical study on colorectal cancer in Mashhad, Iran, and finally predicting the colorectal location of cancer based on the clinical data by using data mining science and decision tree model. Materials and Methods: The data of 316 patients with colorectal cancer (including 14 features) were extracted from the archive of Imam Reza Hospital, Mashhad. The instrument used in this research was RapidMiner data mining software that would try to be extract the details of the relevant data by statistical surveys and then would do initial simulations and the use of classification and decision tree method have predicteion the location of cancer. Results: Male to female ratio of 56% to 44%, family history of 37%, more young patients, and relatively more distally located cancers (39%) compared with the proximal (35%), and rectum (26%) were the striking findings of this study. The final and most important stage of research models were presented, which was able to predict the location of the cancerous tumor with 80% accuracy. Conclusion: Similarities with global statistics, such as the ratio of men to women and family history were observed. But there were also differences with global statistics including the Iran's younger patients and relatively more patients with distal cancers. The efficiency of data mining techniques to predict the location of cancer as well as cost reduction was among the most important results of this study. [ABSTRACT FROM AUTHOR]

Change the world through information technology and Internet development, has created competitive knowledge in the field of electronic commerce, lead to increasing in competitive potential among organizations. In this condition The increasing rate of commercial deals developing guaranteed with speed and light quality is due to provide dynamic system of electronic banking until by using modern technology to facilitate electronic business process. Internet banking is enumerate as a potential opportunity the fundamental pillars and determinates of e-banking that in cyber space has been faced with various obstacles and threats. One of this challenge is complete uncertainty in security guarantee of financial transactions also exist of suspicious and unusual behavior with mail fraud for financial abuse. Now various systems because of intelligence mechanical methods and data mining technique has been designed for fraud detection in users' behaviors and applied in various industrial such as insurance, medicine and banking. Main of article has been recognizing of unusual users behaviors in e-banking system. Therefore, detection behavior user and categories of emerged patterns to paper the conditions for predicting unauthorized penetration and detection of suspicious behavior. Since detection behavior user in internet system has been uncertainty and records of transactions can be useful to understand these movement and therefore among machine method, decision tree technique is considered common tool for classification and prediction, therefore in this research at first has determinate banking effective variable and weight of everything in internet behaviors production and in continuation combining of various behaviors manner draw out such as the model of inductive rules to provide ability recognizing of different behaviors. At least trend of four algorithm Chaid, ex_Chaid, C4.5, C5.0 has compared and evaluated for classification and detection of exist models on the real transactions of private bank. [ABSTRACT FROM AUTHOR]

Introduction: Using data mining to develop prediction models from large amounts of data in medicine has become increasingly popular during the recent years. In this study, we intend to use a decision tree data mining algorithm in order to identify factors influencing mortality in burn patients. Methods: The present retrospective descriptive study is based on burning patients’ records. Overall, the medical records of 4804 patients were scrutinized. The collected data were analyzed using statistical software (SPSS version 16), data mining software (Clementine version 12), and CHAID algorithm. Results: The resulting model for predicting survival and mortality of burning patients included the percentage of Total Burn Surface Area (TBSA %), degree of burn, length of stay, gender, and age of patients. Other variables including blood cultures, wound cultures, urine cultures, and the months of patient hospitalization had no effect on improving the efficiency of the model. Conclusion: The proposed model is valid and reliable due to its accuracy (approximately 95%). In fact, the results of this study, some of which are consistent with the results of other studies, can propose a comprehensive, accurate, and reliable local model for predicting mortality and survival of burning patients based on the mentioned variables. Thus, this local model can be used as an important criterion for evaluating the effective treatment of burn patients. [ABSTRACT FROM AUTHOR]

increasing rate of about 6%. More than 2 million people in Iran are now affected by this disease. The present research deals with the relation between the observed complications of type 2 diabetic patients and some related features like Blood Glucose Level, Blood Pressure, Age, and Family History. The main purpose was to predict the patients’ complications based on the observed signs. Methods: The research data were gathered from 856 patient records related to the 2009’s cases in the Diabetes Center of Golestan province. A new model based on the standard methodology CRISP was developed. In the modeling section, two well-known data mining techniques called C5.0 decision tree and Neural Network were used. Celementine 12.0 software was implemented For data analysis. Results: The results of data mining showed that the variables of high blood pressure, age, and family history had the most impact on the observed complications. Based on the created decision tree, some rules have been extracted which can be used as a pattern to predict the probability of occurring these complications in the patients. The accuracy of the C5.0 model on the data was shown to be 89.74% and on the Artificial Neural Network was 51.28%. Conclusion: As the highest accuracy was shown to be achieved using C5.0 algorithm, according to the created rules, it can be predicted which complication(s) any diabetic patient with new specified features may probably suffer from. [ABSTRACT FROM AUTHOR]

The main goal of this study is the prediction of companies listed in Tehran Stock Exchange using decision tree and logistic regression, which are data mining methods and can help facilitate prediction. Financial ratios are independent variables and healthy and bankrupt companies are dependent variables. Statistical population of the research is information about financial statements of companies listed in Tehran Stock Exchange during 1999 - 2010. No sampling was used in this study; we have two groups of healthy and bankrupt companies. Bankrupt companies group was selected based on article 141 of Commercial law and the healthy group was selected based on profitability criteria. Results suggest that both decision tree and logistic regression methods predict bankruptcy with different accuracy. Area under ROC curve in logistic regression model is more than in decision tree model. [ABSTRACT FROM AUTHOR]

Background: Data mining is known as a process of discovering and analysing large amounts of data in order to find meaningful rules and trends. In healthcare, data mining offers numerous opportunities to study the unknown patterns in a data set. These patterns can be used to diagnosis, prognosis and treatment of patients by physicians. The main objective of this study was to predict the level of serum ferritin in women with anemia and to specify the basic predictive factors of iron deficiency anemia using data mining techniques. Methods: In this research 690 patients and 22 variables have been studied in women population with anemia. These data include 11 laboratories and 11 clinical variables of patients related to the patients who have referred to the laboratory of Imam Hossein and Shohada-E- Haft Tir hospitals from April 2013 to April 2014. Decision tree technique has been used to build the model. Results: The accuracy of the decision tree with all the variables is 75%. Different combinations of variables were examined in order to determine the best model to predict. Regarding the optimum obtained model of the decision tree, the RBC, MCH, MCHC, gastrointestinal cancer and gastrointestinal ulcer were identified as the most important predictive factors. The results indicate if the values of MCV, MCHC and MCH variables are normal and the value of RBC variable is lower than normal limitation, it is diagnosed that the patient is likely 90% iron deficiency anemia. Conclusion: Regarding the simplicity and the low cost of the complete blood count examination, the model of decision tree was taken into consideration to diagnose iron deficiency anemia in patients. Also the impact of new factors such as gastrointestinal hemorrhoids, gastrointestinal surgeries, different gastrointestinal diseases and gastrointestinal ulcers are considered in this paper while the previous studies have been limited only to assess laboratory variables. The rules of the decision tree model can improve the process of diagnosing and treatment of the patients with iron deficiency anemia and reduce their costs. [ABSTRACT FROM AUTHOR]