Modelling reasons for reading problems experienced by third graders through artificial intelligence method

Reading is one of the important indicators of qualified education through which one acquires basic knowledge and gains life-long learning skills. Problems with reading, considered as an essential skill for almost all learning, may affect individuals throughout their lives and prevent them from accessing to qualified education. Reading problems experienced by children are reported as the most common reason why they are diagnosed with learning disabilities. It is also known that some students with no learning difficulties still experience reading problems. Thus, this study inquires the causes of the reading problems experienced by children, who are not diagnosed with learning disabilities, do not have any mental or physical problems but still experience reading problems, based on teacher opinions. In this respect, the study presents five main sections as follows: "the problems experienced by students in reading" and the underlying "familial factors", "individual factors", "social and environmental factors" and "problems caused by educational practices" that cause these problems. This study which draws on the exploratory sequential design, was performed in two steps. In the first step, the qualitative part of the study explored the reading problems that the students experienced as well as the factors underlying these problems under five sections based on the opinions of a total of 47 classroom teachers who participated in this study. A checklist was formed based on the data obtained in this first step. In the second step, the quantitative part of the study, the checklist was administered to a total of 174 classroom teachers teaching the 3rd graders at the primary school level. The data obtained were then converted into numerical values. These values were processed into MATLAB, an artificial intelligence program, and ten different models were developed for the reading problems experienced by the students and the causes of these problems by using artificial neural networks. As a result, a total of 127 items were identified in the qualitative part of this study. According to this, 23 of these items were grouped under "the problems experienced by students in reading", 26 items under "the familial factors underlying the problems experienced by students in reading", 24 items "the individual factors underlying the problems experienced by students in reading", 19 items "the social and environmental factors underlying the problems experienced by students in reading" and 35 items "the problems caused by educational practices underlying the problems experienced by students in reading." The quantitative part was performed based on the checklist formed. The answers given by the teachers in the checklist were converted into numerical values and processed into MATLAB. Through artificial neural networks, two different models as prediction and classification were created for each category and a total of 10 different models were created across all categories. All of these prediction and classification models were developed with one-hundred percent accuracy and near-zero error rates. In conclusion, this study aims to identify the factors for the causes of the possible reading problems that may be experienced by students. Also, it examines the reasons for such problems and contributes to developing solutions to these problems before they arise.