Search

Missing or unavailable data (NA) in multivariate data analysis is often treated with imputation methods and, in some cases, records containing NA are eliminated, leading to the loss of information. This paper addresses the problem of NA in multiple factor analysis (MFA) without resorting to eliminating records or using imputation techniques. For this purpose, the nonlinear iterative partial least squares (NIPALS) algorithm is proposed based on the principle of available data. NIPALS presents a good alternative when data imputation is not feasible. Our proposed method is called MFA-NIPALS and, based on simulation scenarios, we recommend its use until 15% of NAs of total observations. A case of groups of quantitative variables is studied and the proposed NIPALS algorithm is compared with the regularized iterative MFA algorithm for several percentages of NA. [ABSTRACT FROM AUTHOR]

Multiple factor analysis was used to examine organoleptic coffee assessments such as aroma, aftertaste, flavor, acidity, balance, body, uniformity, sweetness, clean cup, and other organoleptic-related properties used in Coffee Quality Assessment. The Sensory analysis was performed using missing values (NA) scenarios with 5%, 10%, 20%, and 30% of NA. The results suggest that RI-MFA is robust to NA presence of and appears to be appropriate when sensory data are present. Simulation scenarios deleting or replacing values from real-world datasets could be a good strategy; different domains, samples, types of variables, and distributions could prove much closer to reality. [ABSTRACT FROM AUTHOR]

In this paper we propose to use the principle of available data derived from the NIPALS (Nonlinear estimation by Iterative Partial Least Square) algorithm to work on the Principal Components Analysis (PCA) in the presence of missing data. This proposal is important since it does not perform data imputation, nor are individuals discarded from the database; the proposed method works with the available pairs to form the quasicorrelation matrices in 𝑹𝒑 and in 𝑹𝒏; the spectral decomposition of these matrices allows through the transition relations to realize a conventional PCA. From the simulation study carried out, it was found that as the percentage of missing data increases, the inertia explained in the first factorial plane decreases. The solution algorithm was developed under the R programming environment and the code is appended for free use. [ABSTRACT FROM AUTHOR]