Principal Component Analysis (PCA) is a multivariate statistical technique to reduce the volume of the large dataset with minimum loss of information and identified the major components along with which maximum variance associated. Generally, PCA was performed using varimax rotation with Kaiser Normalization. The output of the analysis is expressed in terms of Principal Components, their % variance, and factor loading of the variable. In practice, the values of factor loading greater than 0.6 are considered as significant variables. Many times, the values of factor loading are lesser than 0.4. What does it mean? How the result to be interpreted in this case having lower factor loading?
It means that the elements with lower factor loadings are only weakly (or non-) related to the underlying factor. Whether or not you want to consider them as not demonstrated to be related is a matter of personal commitment to significance testing, as well as theoretical considerations.
Dear Sunil,
I wonder where did you get this 0.4 from? I've found the statements saying that if factor loading is close to zero its impact is weak. However, its maximum is at 1.
I wonder also why do you dig factor loading but not the % of variance explained by the reduced number of types of data? As the aim is to reduce it, the maximum variance - after the reduction - should be analysed the most.
Best regards
Hubert
Dear Hubert,
You are absolutely right about factor loading i.e. 0 for weak and maximum of 1. After PCA, the factor loading to components varies from 0 to 1. Particularly the analysis performed for a database, the highest factor loading was found to be 0.4. Is it possible to get many factors having loading more than 0.8 ?
- Badges
- Science topic
- Similar topics
- Philosophy
- Esthetics
- Interpretation