Interventions and schemes aimed at reducing the risks associated with pregnancy are often implemented wholesale. This often leads to misdirection of interventions to the wrong patients. To address this, there was the need to extract segments that subgroup maternal attributes into frequently occurring patterns. Some secondary data consisting of records of pregnant women who attended antenatal care (ANC) visits at a hospital were subjected to association rules mining. The analysis extracted and sub-grouped the attributes and characteristics of pregnant women that often co-occur. Segmentation was done in three sub-groups. Each segment consists of both positive and risk attributes/characteristics that often occur together in a pregnant woman. With the aid of the segments, intervention programs can be designed and delivered according to the needs and requirements of each segment. This provokes a new perspective on how quality healthcare service delivery can be channeled to pregnant women based on specific needs.

Association rules mining technique was employed to extract 6 rules that show the co-occurrences of the attributes on illicit drug types, suspects’ demographics, and categories of drug offences. A dataset on 262 arrestees of various drug offences was utilized for rules extraction using the apriori algorithm. The rules reveal the different levels of involvement with various illicit drugs by suspects of varying ages. The established rules provide a form of drug suspects segmentation which could guide how drug control and intervention programs are designed and deployed. Further, the rules could serve as a reference tool for security agents when dealing with drug suspects and offenders.

Studies on variable selection in the medical field have focused largely on algorithms with little attention paid to domain experts in this regard. This chapter compared the performance of domain experts with filter algorithms in variable selection for clinical predictive modeling. Five clinical datasets on bacterial survival, neonatal birthweight, breast cancer, diabetes, and myocardial infarction were employed. For each dataset, fifteen domain experts were requested to rank the importance of the variables on a five-point Likert scale. The same variables were ranked using four algorithms, namely, chi-squared, Fisher score, Pearson's correlation, and varImp function. Results of classification models showed that both methods performed competitively. This means human expertise and experience are important in clinical predictive modeling and must not be mortgaged to algorithms. Further studies should focus on developing automated platforms that codify domain knowledge and experience to facilitate real-time, speedy, and seamless variable selection.

No abstract