I have working with heavy metals to reduce the data set i used to make a PCA with the help of PAST tool. Right now i got all those things like score plot and all.. Finally how can i interpretation the output?
Hello GaneshKumar:
PCA gives new indicators which are linear combinations of the original ones, thus the new indicators combines similar old indicators through their shared properties, you are going to redefine these new indicators according to your understanding of the potential shared properties. You are also going to choose a proper number of new indicators according to how much information is interpreted by these new indicators. Through the process, the number of indicators is reduced.
I am not sure what the score plots are, because I use other platform to perform PCA, but the main idea is that the results may indicate (1) how the new indicator is composed of the original one, and (2) how the new indicators interpret the information through variance or eignenvalues.
Regards,
Jiong
Observing the weightage value of parameters, but there will be noise in each value
Hi,
Theoretically, PCA is a method of creating new variables (known as principal components, PCs), which are linear composites of the original variables. The values of PCs created by PCA are known as principal component scores (PCS). The maximum number of new variables is equivalent to the number of original variables.
To interpret the PCA result, first of all, you must explain the scree plot. From the scree plot, you can get the eigenvalue & %cumulative of your data. The eigenvalue which >1 will be used for rotation due to sometimes, the PCs produced by PCA are not interpreted well. Consequently, the varimax rotation has been applied to rotate the PCs for the interpretation purposes. Eigenvalues obtained from varimax rotation are the precursor of PCA. Eigenvalues >1.0 were considered as significant and subsequently varimax factors (VFs), which are the new groups of variables are generated. The VFs values which are greater than 0.75 (> 0.75) is considered as “strong”, the values range from 0.50-0.75 (0.50 ≥ factor loading ≥ 0.75) is considered as “moderate”, and the values range from 0.30-0.49 (0.30 ≥ factor loading ≥ 0.49) is considered as “weak” factor loadings.
To me, only VFs value >0.75 are considered for selection and interpretation due to having significant factor loadings. From the highest value (>0.75) of VFs, then you can reduce the parameter without reduce dataset. The interpretation of your output is actually based on what you want to put into your paper.
Regards,
Azman Azid
My suggestion: https://www.researchgate.net/publication/319469038_New_Interpretation_of_Principal_Components_Analysis
Article New Interpretation of Principal Components Analysis
i really want to know how to interpret Rotated factor loadings (pattern matrix) and unique variances, Bartlett test of Sphericity and Kaiser-Meyer-Olkin. I would appreciate some examples with atleast 5 to 6 factors
PCA is a multivariate test that aim to consize the uncorrelated variables as principle components. These loading are expressed as principal components. The graphical representation is expressed as PCA.
see the number of PC components that explained around 80 to 90 % variance and used those components only further in your model...
This can be useful https://www.reneshbedre.com/blog/principal-component-analysis.html
Excellent interpretation is made in the article Article New Interpretation of Principal Components Analysis
This will help to grasp in-depth understanding
in PCA, two points are very important:
1. Eigenvalues: It is used to represent the importance of features
2. Cumulative % variance: It is used to represent the gives the percentage of variance accounted for by the first n components
Please go through this paper:
Hybrid functional link artificial neural network
approach for predicting maintainability of
object-oriented software
in the interpretation of the principal component analysis (PCA), you can decide to identify the items associated with the highest loaded score for each component which is highly uncorrelated with other components' scores, and subject them for further analysis as the most significant representatives of the latent construct with the highest variance .
Secondly, depending on the type of software you are using for the analysis, you can generate a factor score series from the principal component analysis which serves as the series for the latent construct being measured if you are interested in conducting a further investigation with the latent factor.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis