Clustering MBTI Personalities With Graph Filters And Self Organizing Maps Over Pinecone

Sep 17, 2026

This conference paper presents a methodology for clustering Myers-Briggs Type Indicator (MBTI) personalities using self organizing maps (SOMs) enhanced by graph filters. Topologically flexible distance metrics are constructed by combining tensors with graph signal processing operators, enabling higher-order relationship modeling within dense attribute spaces. Multiple graph filters are evaluated on a Kaggle benchmark with ground truth using topological error, cluster purity, inter-cluster distance, and curvature variability. Data are stored in Pinecone and accessed via Python integration.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages8

Reading Time20 min

AccessFree to download

Abstract

Introduction

Previous Work

Clustering MBTI Personalities With Graph Filters And Self Organizing Maps Over Pinecone Georgios Drakopoulos Department of Informatics Ionian University 0000-0002-0975-1877 Phivos Mylonas ICE Department University of West Attica 0000-0002-6916-3129 Abstract—Self organizing maps (SOMs) or cognitive maps are designed to preserve major topological attributes of manifolds ina higher dimensionality data space to corresponding projections thereof in a low dimensional coordinate space. This is performed by mapping neighborhoods and the distances contained therein from the data space to ones in the coordinate space. Thus, SOM functionality relies heavily on the geometrical properties of both spaces. Topologically ﬂexible data space distance metrics are constructed by combining tensors with graph ﬁlters, the latter coming from graph signal processing. The power of these distance metrics comes from naturally expressing the higher order relationships between points of the data space. This paves the way for addressing engineering scenarios involving a large number of densely interrelated attributes. One such case is discerning personalities from textual information based on the Myers-Briggs taxonomy indicator (MBTI), a framework of archetypal personalities derived from Jungian psychodynamic theory. Various graph ﬁlters were tested on a benchmark Kaggle dataset with ground truth with comparisons assessed in terms of topological error, cluster purity, average inter-cluster distance, and cluster curvature variability. Data points were stored in Pinecone, a recent vector database, with Python integration. Index Terms—SOMs, graph ﬁlters, graph signal processing, graph topology, tensor distance metrics, geometric analytics, cluster curvature, dimensionality reduction, MBTI, Pinecone I. INTRODUCTION Self organizing maps (SOMs) or cognitive maps are unsupervised neural network architectures designed to approximate a manifold in a high dimensional data space V by constructing an approximately continuous projection thereof, a topological map, in a lower dimensional coordinate space C [1] . Geometry is crucial in shaping said projection as the latter preserves up to a point the primary topological properties of the original manifold [2] . Dimensionality reduction is thus achieved, which may be an intermediate step for clustering or visualization. Human personality and the evolving dynamics underpinning their interplay have been for quite a number of years central research points across domains such as psychology [3], social network analysis [4], online marketing and digital campaign design [5], computational neurosciences [6], and sociology [7] . Gaining insight into team dynamics allows informed decisions in tasks so diverse as workgroup formulation, mentor matching, designing unobviated and unobtrusive communication, and task delegation. The basic building block of any TABLE I NOTATION SYNOPSIS . | Symbol | Meaning | First in | | --- | --- | --- | | △
=
{s1 , . . . , sn}|·|
loc (·) invloc (·) weight (u)Γ (u)
∆ (u; η 0)⋆
∥·∥Ff (k) (x) prob {Ω} | Deﬁnition or equality by deﬁnition | Eq. (1) | | | Set with elements s1 , . . . , sn | Eq. (1) | | | Set cardinality | Eq. (14) | | | Location function for data points | Eq. (1) | | | Inverse location relationship for neurons | Eq. (1) | | | Synaptic weights of neuron u | Eq. (3) | | | Neighborhood of neuron u | Eq. (2) | | | Cover of neuron u with threshold η 0 | Eq. (2) | | | Application of spatial ﬁlter to graph | Eq. (7) | | | Frobenius tensor norm | Eq. (7) | | | k-th derivative of function f | Eq. (23) | | | Probability of event Ω occurring | Eq. (16) | such approach is a methodological and systematic framework for characterizing human personalities based on fundamental psychological traits [8] . One such model is the Myers-Briggs Type Indicator (MBTI) which is derived from the Jungian psychoanalytical framework [9] . The latter is a personality taxono

FAQs

What role do self organizing maps play in the proposed approach?

They are used to construct a topological projection of high-dimensional data into a lower-dimensional space suitable for clustering and analysis.

How are graph filters incorporated into the SOM-based method?

Graph filters derived from graph signal processing are combined with tensors to build topologically flexible distance metrics that capture higher-order relationships among data points.

How is the method evaluated and where is the data stored?

Filters are compared on a Kaggle dataset using metrics such as topological error, cluster purity, average inter-cluster distance, and cluster curvature variability, while vector data are stored in Pinecone with Python integration.

Sep 17, 2026

Uploaded by Genevieve