In the context of adoption, evaluating prospective adoptive parents using psychometric assessments such as the Minnesota Multiphasic Personality Inventory (MMPI) questionnaire is essential for understanding their psychological profiles. However, interpreting such complex data can be both challenging and time-consuming. In this study, we propose a meta-analysis tool to assist psychologists in their initial interpretation and analysis of MMPI-2 results by providing a clear data-driven visualization of key psychometric scales. Our system employs unsupervised learning techniques to uncover meaningful patterns and relationships in the data with minimal prior input. Specifically, a genetic algorithm is used to optimize clustering quality by selecting the most relevant psychological scales, enhancing cluster separation, and improving data interpretability. We also explored and compared the efectiveness of several clustering algorithms, including K-Means, Gaussian Mixture Model, and Spectral Clustering, to maximize the capabilities of our tool.
M
MPI-2 and Evolutionary
Adoption is the process whereby individuals or families
assume the parenting of a child who is not biologically
their own. According to specific studies [
released MMPI-3 [
For the evaluation of the results, the set of most important psychometric scales to be analyzed is usually handpicked by field experts as it is highly task-dependent. For that reason, in this study, we propose an unsupervised learning algorithm capable of clustering the data gathered with the MMPI-2 test using as little as possible prior knowledge during the preprocessing and postprocessing of the data.
ogy traits of prospective adoptive parents. An example of 7, 8] are used to assess the personality and psychopathol- technique designed to identify similarities within data without predefined categories. In our case, by analyzing such a test is the Minnesota Multiphasic Personality In- the geometric properties of the data, the goal is to capture CEUR
ceur-ws.org
specifically for adults; the MMPI-A [
developed. The most commonly used versions today
include the MMPI-2 [
(S. Russo) (S. Russo) ICYRIME 2024: 9th International Conference of Yearly Reports on [28]) to determine the most suitable one for our system.
of the MMPI-2 questionnaire, its scales, and traditional Following this concise overview of the MMPI-2 test, prior
MMPI clustering methods is presented in Section 2. Next, attempts to cluster datasets derived from this assessment
Section 3 provides a detailed description of the core tech- have typically involved manually selecting sets of the
niques used in our algorithm. In Section 4, we describe aforementioned psychometric scales.
the dataset employed in this experiment. Following this, In [29], an algorithm very similar to K-Means
(origSection 5 ofers a comprehensive explanation of the sys- inally described in [30]) was applied to data obtained
tem we developed and its evaluation process. The clus- from MMPI-2 tests administered to women in their third
tering results produced by our system are then presented trimester of pregnancy. The objective was to determine
in Section 6. Finally, Section 7 summarizes the article’s the personality characteristics of women who develop
content and outlines potential areas for future improve- perinatal depression.
ment. Similarly, in [
Validity scales scrutinize the subject’s approach to membership, but there were discrepancies in the reported
the test and demeanor, identifying inconsistencies or clarity of PTSD education materials among clusters.
attempts to manipulate responses. Among them, the Lie In [
In addition, the MMPI incorporates ten primary clini- tion feature in IBM SPSS 21 to select the optimal cluster cal scales designed to detect a spectrum of psychological solution. Cluster analysis revealed two distinct patient disorders, encompassing Hypochondriasis (Hs), Depres- clusters. Cluster 1 displayed valid responses and exhibsion (D), Hysteria (Hy), Psychopathic Deviate (Pd), Mas- ited elevations primarily on somatic and low positive culinity/Femininity (Mf), Paranoia (Pa), Psychasthenia emotion scales. In contrast, Cluster 2 comprised patients (Pt), Schizophrenia (Sc), Hypomania (Ma), and Social who overreported on validity scales and demonstrated Introversion (Si). Furthermore, content scales target spe- elevations on multiple restructured clinical scales. cific personal attitudes, including anger issues (ANG), low self-esteem (LSE), family problems (FAM), and workrelated challenges (WRK), among others. 3. Core Techniques in Our
Additionally, supplemental scales are used in combina- Algorithm tion with the content scales to determine if some symptoms are attributed to alternative potential causes such 3.1. Genetic Algorithm as controlled hostility, alcoholism, and more.
Moreover, Psy-5 scales measure dimensional traits of personality disorders, including Aggressiveness, Psychoticism, Constraint, Neuroticism, and Extraversion.
Finally, to ensure uniform interpretation across all scales, scores are transformed into T-scores, ranging from 30 to 120. Typically, scores exceeding 65 are considered significant and warrant further examination.
All cited works in this paper employ clustering
techniques with input from psychology experts to select
relevant psychometric scales for analysis. In contrast, our
system autonomously selects key scales using a genetic
algorithm [
They involve successive generations of candidate
solutions undergoing reproduction, mutation, and selection ifned number of clusters through an iterative process:
to converge toward optimal or near-optimal solutions. randomly selecting K samples as initial clusters (and
cenGenetic algorithms have a broad range of applications troids), assigning each sample to the cluster with the
[
In summary, a general genetic algorithm workflow is below a certain threshold.
the following: firstly, an initial population of individuals Gaussian Mixture Model (GMM) endeavors to fit a
(each represented as a string of 0s and 1s) is randomly specified number (N) of normal distributions to
disgenerated. Next, a fitness value is assigned to each in- tinct subsets of the original dataset by estimating their
dividual in the population according to a certain fitness mean and variance parameters using the
Expectationfunction. Then, multiple pools of individuals are ran- Maximization (EM) algorithm [
3.2. Clustering Algorithms tries and 813 features for each entry. These features Clustering algorithms belong to the unsupervised learn- encompass anamnestic information, boolean answers to ing domain of artificial intelligence and are designed the MMPI’s questions, and T-scores. Figures 1, 2, and to unveil concealed patterns and organize data points 3 provide an overview of the statistics regarding some into coherent clusters based on their intrinsic similari- of the anamnestic information and the clinical and conties. These algorithms rely on diferent distance metrics tent scales, calculated as T-scores, of the subjects in our like Euclidean distance, cosine similarity, and the Jac- dataset. For preprocessing, we removed features with card coeficient to quantify the resemblance between either a single value or a predominant value (e.g., ‘Citizendata points. The typical representation of each resulting ship’) and those with high variability (e.g., ’Profession’). cluster involves a centroid, acting as a central reference Additionally, we dropped the gender column since MMPI point summarizing the collective traits of its constituent scales have the same interpretation for both men and data points. These algorithms can be broadly categorized women. The boolean answers to the MMPI’s questions into several methodologies. Partitioning methods, exem- were also discarded, as the normalized T-score values plified by K-means, iteratively segment the dataset into automatically encode this information. non-overlapping clusters, ensuring each data point exclu- To ensure data validity, according to the guidelines sively belongs to one cluster. Hierarchical methods, such provided by the authors of the MMPI test, applicants as Agglomerative clustering, construct a hierarchical ar- with Lie scale scores exceeding 75 were excluded. Adrangement of clusters by iteratively merging or dividing ditionally, none of the test-takers reached the cutof of existing clusters based on similarity criteria, culminating 30 unanswered questions on the ’cannot say’ scale that in a tree-like structure. Model-based methods, on the should invalidate the test. We also examined other vaother hand, assume that the data is generated by a proba- lidity scales such as F, TRINT, and VRINT, but no enbilistic model, such as a Gaussian Mixture Model (GMM), tries were excluded based on these scales. Applicants allowing for the probabilistic modeling of clusters. with high values indicating alcohol or drug issues were
In our study, we focus on evaluating and comparing marked as rejected in advance.
the performance of K-means, Gaussian Mixture Model, The remaining data, consisting of 191 entries with 120
and Spectral Clustering. feature columns, was scaled to ensure all features had
In detail, K-Means partitions samples into a prede- the same magnitude within the range [
feature in our dataset has been encoded with a binary
digit [
where is the mean distance between a sample and all other points in the same cluster, and is the mean distance between a sample and all points in the nearest cluster. The Silhouette Score, which is the average of the Silhouette Coeficients for all elements in the dataset, indicates the quality of clustering. A higher mean Silhouette Score suggests denser and better-separated clusters. In our study, the optimal number of clusters found for our dataset was 2, as shown in Fig. 6. The Fig. 7 provides a comprehensive overview of Silhouette coeficients for diferent numbers of clusters, demonstrating the decline in clustering quality as the number of clusters increases.
Executing the PCA to the obtained clusters generates the plot displayed in Fig. 8. It can be seen that on the ifrst principal component (x-axis) the two clusters are well distinguished while on the second principal component (y-axis) they both spread homogeneously even if the elements belonging to the green cluster are more concentrated around the zero value of that axis.
In a more detailed analysis, Fig. 9 illustrates the intracluster average values for the four main group scales: Validity, Clinical, Content, and Supplemental. As observed, the elements in the green cluster consistently show lower average values compared to those in the red cluster, with the exception of the Validity scale. This reversal in trend may prompt psychologists to further examine these two clusters, as the scales within the Validity group are designed to indicate how reliable and truthful the test responses are. However, the diferences between the clusters are minor, and both demonstrate a high level of reliability in responses, with few outliers. One of the key insights from this analysis is the notable diference in the Content scale, suggesting that individuals in the red cluster may exhibit more psychological issues compared (3) to those in the green cluster.
A similar trend, observed in Fig. 8, is also highlighted employed the Silhouette Analysis. This technique involves computing the Silhouette Coeficient for each element in the dataset, defined by: =
− max(, ) in Fig. 11, where the x-axis represents the average values of the Content scale and the y-axis the average values of the Clinical scale for each element in the dataset.
Finally, Fig. 10 provides a deeper analysis of the weights associated with the psychological scales for the ifrst and second principal components of the PCA. From this plot, it is clear that for the elements in the green cluster, high values on scales related to the Content group correspond to highly positive weights, while low values correspond to negative weights. In contrast, the red cluster exhibits an inverted trend. For the second principal component, the red cluster elements are more evenly distributed across the dimension, while the green cluster elements generally show lower values across the scales. From these graphs, psychology experts can gain insights into the most relevant psychological scales within the
MMPI-2 profiles of prospective adoptive parents using evolutionary clustering techniques. By incorporating a genetic algorithm to autonomously select the most relevant psychometric scales, we aimed to streamline the clustering process and reduce reliance on manual selection by domain experts.
By employing a genetic algorithm to automatically select the most relevant psychological scales, combined with K-Means clustering based on minimum centroid distance and Silhouette analysis, we determined that two clusters were the optimal choice to describe the analyzed dataset.
These clusters displayed distinct psychological profiles, with notable diferences particularly in the content and clinical scales, which may serve as valuable insights for psychologists when assessing potential adopters.
The implications of our approach are twofold: first, it ofers a data-driven methodology that enhances the initial interpretation of complex MMPI-2 profiles, assisting psychologists in identifying meaningful patterns without prior assumptions. Second, it underscores the potential of unsupervised learning techniques, such as genetic algorithms, in improving psychometric data analysis by automating feature selection and optimizing clustering quality.
Future work may involve expanding the dataset and further refining the genetic algorithm to handle larger and more diverse MMPI profiles. Additionally, exploring the integration of other clustering methods and incorporating newer versions of the MMPI test, such as MMPI-3, may provide further improvements and adaptability in diverse psychological evaluations.