Let quantita represent an unknown document and let y represent per random target author’s stylistic ‘profile’. During one hundred iterations, it will randomly select (a) fifty verso cent of the available stylistic features available (e.g. word frequencies) and (b) thirty distractor authors, or ‘impostors’ from per pool of similar texts. Mediante each iteration, the GI will compute whether quantita is closer puro y than preciso any of the profiles by the thirty impostors, given the random selection of stylistic features per that iteration. Instead of basing the verification of the direct (first-order) distance between incognita and y, the GI proposes puro record the proportion of iterations in which quantita was indeed closer onesto y than to one of the distractors sampled. This proportion can be considered a second-order metric and will automatically be per probability between niente and one, indicating the robustness of the identification of the authors of x https://datingranking.net/it/lumenapp-review/ and y. Our previous rete di emittenti has already demonstrated that the GI system produces excellent verification results for classical Latin prose.31 31 Amico the setup sopra Stover, et al, ‘Computational authorship verification method’ (n. 27, above). Our verification code is publicly available from the following repository: This code is described mediante: M. Kestemont et al. ‘Authenticating the writings’ (n. 29, above).
For modern documents, Koppel and Winter were even able puro report encouraging scores for document sizes as small as 500 words
We have applied per generic implementation of the GI sicuro the HA as follows: we split the individual lives into consecutive samples of 1000 words (i.ancora. space-free strings of alphabetic characters), after removing all punctuation.32 32 Previous research (see the publications mentioned in the previous two taccuino) suggests that 1,000 words is a reasonable document size per this context. Each of these samples was analysed individually by pairing it with the profile of one of the HA’s six alleged authors, including the profile consisting of the rest of the samples from its own text. We represented the sample (the ‘anonymous’ document) by verso vector comprising the imparfaite frequencies of the 10,000 most frequent tokens in the entire HA. For each author’s profile, we did the same, although the profile’s vector comprises the average incomplete frequency of the 10,000 words. Thus, the profiles would be the so-called ‘mean centroid’ of all individual document vectors for a particular author (excluding, of course, the current anonymous document).33 33 Koppel and Seidman, ‘Automatically identifying’ (n. 30, above). Note that the use of per celibe centroid a author aims to ritornato, at least partially, the skewed nature of our data, since some authors are much more strongly represented mediante the corpus or preparazione pool than others. If we were not using centroids but mere text segments, they would have been automaticallysampled more frequently than others during the imposter bootstrapping.
To the left, verso clustering has been added on sommita of the rows, reflecting which groups of samples behave similarly
Next, we ran the verification approach. During one hundred iterations, we would randomly select 5,000 of the available word frequencies. We would also randomly sample thirty impostors from per large ‘impostor pool’ of documents by Latin authors, including historical writers such as Suetonius and Livy.34 34 See Appendix 2 for the authors sampled. The pool of impostor texts can be inspected con the code repository for this paper. Sopra each iteration, we would check whether the anonymous document was closer puro the current author’s profile than puro any of the impostors sampled. Con this study, we use the ‘minmax’ metric, which was recently introduced durante the context of the GI framework.35 35 See Koppel and Winter, ‘Determining if two documents’ (n. 26, above). For each combination of an anonymous text and one of the six target authors’ profiles, we would superiorita the proportion of iterations (i.anche. a probability between zero and one) sopra which the anonymous document would indeed be attributed to the target author. The resulting probability table is given sopra full con the appendix esatto this paper. Although we present a more detailed tete-a-tete of this momento below, we have added Figure 1 below as an intuitive visualization of the overall results of this approach. This is a heatmap visualisation of the result of the GI algorithm for 1,000 word samples from the lives mediante the HA. Cell values (darker colours mean higher values) represent the probability of each sample being attributed onesto one of the alleged HA authors, rather than an imposter from per random selection of distractors.