Voynich : hoax?

Abstract

The Voynich Manuscript, known as ‘the most mysterious manuscript in the world’ [1] [2] has undergone tremendous efforts to decode its mysterious text which, so far, has been unsuccessful. The manuscript is believed to be divided into six major categories based upon the illustrations on the pages which act as indicators of the probable topics. These topics are: Herbal, Astronomical, Biological, Cosmological, Pharmaceutical and Recipes [1]. The goal of this research is to examine whether the Voynich manuscript is an elaborate hoax or an unknown language. If the manuscript is the ‘real deal’, as suggested by Reddy and Knight [3] - namely that the Voynich is written in a real, but unknown language - then we should expect that, on average, text within different categories are more semantically related to each other than text between different categories. To test this hypothesis a statistical approach is taken and applied to the Voynich Manuscript. This research is performed using Extensible Voynich Alphabet (EVA) transcription which is digitally available to the public [4]. Since this dataset is an ongoing work and contains terms that have more than one interpretative readings, pre-processing techniques are performed to reduce such ambiguities. Various machine learning algorithms, including classification and clustering techniques, such as Support Vector Machine (SVM), k-Nearest Neighbours (kNN), Multinomial Naïve Bayes, KMeans and No-K-Means are utilised. The strongest results are achieved when each category is extracted based on pages rather than when each category is extracted on a line-by-line basis. Through different metrics and diagrams, these algorithms show clear and expressive results which demonstrate enough evidence that a particular relationship is contained among the categories. This provides support for the existence of a genuine message within the manuscript.