br Characteristics of the database used in

2.2. Characteristics of the database used in this study
The Raman database used comprised a total of 223 Raman spectrum samples obtained in the spectrum range of 400–1800 cm−1 wave-lengths. After final verification, the total number of samples in the Raman spectral database, whose values were obtained from a spectral collection ex vivo, were divided into four histopathological groups. The 223 spectral samples were divided into 115 spectra of 11024-24-1 lesions di-agnosed as BCC, 21 spectra as squamous cell carcinoma (SCC), 57 spectra as actinic keratosis (AK), and 30 spectra of normal skin without lesions (NO).
2.3. Organization of Raman data
From the Raman spectra of each skin type sample (NO, BCC, SCC, and AK), a database was generated; it included 30 normal skin tissues (NO) samples, a group of non-melanoma skin tissues comprising 115 BCC samples and 21 SCC samples, and 57 AK skin tissues samples, as
Table 1
Characteristics of the Raman intensity database with spectral samples of ex vivo skin tissues comprising NO and BCC + SCC and AK in 400–1800 cm−1 spec-trum.
Raman Skin Tissue (NO)
Spectra
Wavelength Sample01 … Sample30
Raman
Skin Tissue (BCC + SCC)
Spectra
Wavelength Sample01 … Sample115
Raman Skin Tissue (AK)
Spectra
Wavelength Sample01 … Sample57
can be seen in Table 1.
Following [19], in which statistical techniques were used to dis-criminate skin lesions in a database of ex vivo-obtained Raman samples of skin tissues, in this work, three distinct types of separations and ar-rangements were used. However, considering that the research is of greater importance to medical diagnosis, the research focus was on the
following arrangement: NO × (BCC + SCC) × AK (normal skin dis-crimination from BCC and SCC, which are considered as a single group, and BCC and SCC discrimination from AK). For the discrimination, all 223 randomly selected samples were analyzed by SPA-PAL2v, aiming to obtain a percentage of correct results for the diagnosis according to this arrangement type.
3. Theory/Calculation
The PAL2v equations were used to form the algorithms of the computational structure of this work. The configuration of the inter-connected algorithms extracted from the fundamentals of PAL2v is called the SPA-PAL2v, a set of paraconsistent algorithms [20–22]. De-tails of the SPA-PAL2v structure are provided below:
3.1. Set of paraconsistent algorithms
The SPA-PAL2v is a set of paraconsistent algorithms based on PAL2v; it forms a computational framework for the analysis of data obtained from Raman spectroscopy of skin cancer tissue samples. The paraconsistent algorithms used in SPA-PAL2v for analysis of the Raman database are as follows:
1 Algorithm for extracting the degrees of evidence [26]
3 Algorithm for extracting the contradiction effects (ParaExtrctr) [27] 4 Algorithm for randomly selecting and extracting the sample evi-
dence degree
5 Algorithm for detecting the number of occurrences of similarity [5] and
6 Algorithm for extracting the evidence degrees of the frequency [5]
The first three algorithms have been detailed in Section 1.5 and in [26,27], while the others are discussed below and applied to the SPA-PAL2v computational structure.
The general form of how SPA-PAL2v is applied in skin cancer di-agnosis is shown in the flowchart of Fig. 3.
As can be seen in Fig. 3, in the first step, the paraconsistent Raman patterns are generated, and in the second step, inquiries are made for the validation of the SPA-PAL2v. The validation is done by verifying the diagnoses correctness through randomly selected samples.
Fig. 3. Flowchart of the set of paraconsistent algorithms (SPA-PAL2v) used to discriminate the skin spectra in one of the groups.
3.2. Algorithm for extracting the degrees of Raman intensity evidence
The algorithm that extracts the degrees of evidence [5] Raman In-tensity (μ) has the action of transforming the values of the spectra of ex vivo skin tissue samples obtained by Raman spectroscopy in evidence degrees ranging from 0 to 1 and belonging to the set of real numbers [4,25]. Initially, the minimum and maximum values of the Raman In-tensity in each line of the wavelength of the diverse samples are es-tablished. Subsequently, the normalization of the Raman intensity va-lues surveyed in the Raman spectroscopy database of each type is characterized by the equation:
1 If Xvalue > Maxvalue
Xvalue − Minvalue If Xvalue ∈ [Min value , Maxvalue]
μPP = Max value − Minvalue