**4. Results**

#### **4.1 Correlation analysis results**

A correlation study was performed, as previously indicated, to check the connection between the input variables and the output power, thereby selecting the closely related factor parameters that should be kept as inputs to the prediction models (see **Figure 1**).

### **4.2 Principal component analysis results**

As previously explained, PCA was used to determine the most essential data variables to be used in the training of the machine learning models. The variance distribution of the principal components (PCs) (PC1–PC9) is depicted in the Scree plot in **Figure 2**. According to the eigenvalues, the cumulative variance of PC1 through PC3 is **90.4**%. As a result, the first three major components were recognized as the primary model inputs and were sufficient for the development of our predictive models.

**Figure 1.** *Correlation matrix.*

**Figure 2.** *Scree plot.*

The main variables of each of the PCs were selected from the top three variables in **Table 4** with a value greater than **0.60** [15]. **GHI**, **BHI**, and **BNI** were selected for PC1. For PC2, **Hour**, **Tm,** and **Eff** were identified. Finally, only **Tamb** was chosen for PC3.

*Principal Component Analysis and Artificial Intelligence Approaches for Solar… DOI: http://dx.doi.org/10.5772/intechopen.102925*


#### **Table 4.**

*PCA results.*


#### **Table 5.**

*Performance metrics results—Training phase 80%.*


#### **Table 6.**

*Performance metrics results—Testing phase 20%.*
