BilalKhan563/Anemia-Factors-ML

Exploring childhood anemia factors using ML on 2018 Nigeria DHS data for insights into prevalence & predictive markers. # Factors Affecting Children Anemia Level This dataset represents a cross-sectional study conducted during the 2018 Nigeria DHS. It aims to measure various factors that potentially influence anemia levels in children aged 0-59 months. ## Overview ### Dataset Description - The dataset comprises several factors that could contribute to anemia levels in children, including demographics, health-related questions, and socioeconomic indicators. - Columns like `Age_in_5year_groups`, `Type_of_place_of_residence`, `Highest_educational_level`, and `Wealth_index_combined` offer insights into potential influences on anemia. ### Data Exploration and Insights - Visualizations revealed intriguing patterns: - **Anemia in Relation to Smoking:** Non-smokers seem more prone to anemia. - **Impact of Recent Fever:** Children who experienced fever in the last two weeks showed a higher likelihood of anemia. - **Medication Intake:** Anemia prevalence seems lower among those taking iron pills, sprinkles, or syrup. ### Model Development and Evaluation - **Data Preprocessing:** Addressed missing values and converted data types for model compatibility. - **Machine Learning Model:** Utilized RandomForestClassifier: - Trained on a split dataset (80% training, 20% testing). - Achieved a model accuracy of 99.38%. ### Model Performance Analysis - While the model showcases high accuracy, further domain-specific considerations might be needed to determine practical implications. - Considerations: - Are false negatives or false positives more critical in this context? - Sensitivity or specificity measures could provide a deeper understanding of the model's predictive power. ### Challenges and Limitations - **Data Completeness:** Missing values, especially in hemoglobin-related columns, were handled by imputation. - **Dataset Scope:** Limitations might exist in capturing all factors influencing anemia, requiring supplementary data or refined modeling techniques. ### Future Steps - **Refinement …