Machine Learning Used to Predict DivorcePublished on Mon Oct 16 2023 by Dustin Van Tate Testa Divorce - Splitting the House | Cathrine Sæther
New research has utilized machine learning algorithms to predict whether a couple is likely to get divorced or stay married. The study, titled "Divorce Prediction with Machine Learning: Insights and LIME Interpretability," explored the use of six different algorithms to accurately classify married and divorced individuals. Surprisingly, the preliminary computational results showed that algorithms such as Support Vector Machines (SVM), K-Nearest Neighbors (KNN), and Linear Discriminant Analysis (LDA) had an impressive accuracy rate of 98.57%.
What sets this research apart is its use of Local Interpretable Model-Agnostic Explanations (LIME) to explain prediction probabilities. By analyzing test results with LIME, researchers were able to differentiate between divorced and married couples effectively. Furthermore, the study identified the ten most crucial factors that potentially contribute to a couple's decision to divorce using top feature selection methods.
The findings of this research are not only valuable for academic purposes but also for individuals who want to assess the health of their own relationships. To this end, the researchers created a user-friendly divorce predictor app based on the identified factors. This tool can be used by anyone to evaluate their relationship condition and take necessary steps to prevent a potential divorce.
While the study's results are certainly promising, it is worth noting that the dataset used in the research was highly balanced. This means that the accuracy achieved during preliminary computations may not be as reliable when applied to real-world scenarios. Future investigations will focus on exploring other machine learning techniques and deep learning approaches to gain a more in-depth understanding of the study's results, particularly with high-dimensional data.
Overall, this research offers hope for individuals seeking ways to predict the likelihood of divorce within their own relationships. By harnessing the power of machine learning algorithms and interpretability techniques like LIME, couples can gain valuable insights into their relationship's future and, if necessary, take proactive measures to maintain a healthy and fulfilling marriage.