Comparative Classification Efficiency of Self-organizing Map, Principal Component Analysis, and Hierarchical Cluster Analysis for Normal Dairy and Differently Pineapple Fortified Rasgulla

Abstract

Rasgulla, the popular Indian dairy-based dessert, was fortified by using pineapple to improve its nutritional and different physical parameters like color, texture and sensory attributes. Being a highly nutritious, palatable fruit with excellent flavor and visual color characteristics, pineapple fortification provides an edge over normal dairy rasgulla. Comparative analysis of nutritional, X-ray diffraction, texture and sensory characteristics for normal rasgulla sample and pineapple with different drying process (pulp (PP), convective (PH), microwave (PMW), microwave convective (PMWC), and freeze-dried (PF)) fortified rasgulla samples were done. PH was found with maximum carbohydrate (40.60±2.04) and minimum protein content (5.91±0.95). PMWC was found with the maximum total color difference (ΔE = 48.79±3.33). Pineapple fortification did not alter the crystal characteristics of the rasgulla samples. Normal rasgulla showed the lowest hardness value while PMWC was the hardest one. The results obtained from the analysis were further comparatively segregated with the help of principal component analysis, hierarchical cluster analysis and self-organizing map to optimize the analysis frequency towards a more accurate result. It was found that more or less the same trends in results were obtained by all three analysis tools. It can be concluded that machine learning-based self-organizing map can be more sophistically used for classification between different parameters even in case of complex non-linear problems also. Among the entire samples microwave, convective dried pineapple rasgullawas found to be the best in terms of all the observed parameters.

The impact of raw and differently dried pineapple (Ananas comosus) fortification on the vitamins, organic acid and carotene profile of dairy rasgulla (sweetened cheese ball)

“…This might be explained as, because of the incorporation of higher moisture from pineapple fruit pulp (80.78 ± 3.2 %) and differently dried pineapple pulp as well. According to different literature, the ash content of normal rasgulla was varied in the range of (0.75 ± 0.005–1.04 ± 0.03) % [ 12 , 13 ] and in this study, the ash content of NR was found 1.19 ± 0.03 %.…”

“…Moisture content (MC) and crude protein of samples were estimated as per AOAC 2000 [ 12 ]. Carbohydrate, fat and ash content were measured according to Sarkar et al., 2020 [ 13 ].…”

“…The protein content of all the pineapple fortified rasgulla samples were significantly lower with respect to NR and this was due to different drying process which caused [19]; Garcia-Amezquita et al, 2018 reported a higher amount of protein content in freeze dried pear-peel compared to hot air drying [20]. NR and pineapple fruit were reported to have protein content (3.91 AE 0.12-8.24 AE 0.05) % and (1.19 AE 0.03) % respectively [12,13,14]. The carbohydrate content of NR (37.02 AE 1.03 %) was found higher than PP (29.616 AE 1.12 %) due to the higher carbohydrate content of NR than raw pineapple fruit pulp.…”

Optimization and modeling of vacuum impregnation of pineapple rings and comparison with osmotic dehydration

“…The rise in YI with an increase in temperature might be due to the degradation of color pigments, carotenoids, and xanthophylls. These color pigments start degrading above 50°C temperature (Sarkar et al., 2020). When sugar in pineapple juice is exposed to high temperatures, it caramelizes, changing the yellowish color of pineapple to an orangish red color (Assawarachan & Noomhorm, 2010).…”

Combination of Computational Techniques to Obtain High-Quality Gelatin-Base Gels from Chicken Feet

“…Sarkar et al [ 32 ] compared the classification efficiency of SOM, PCA, and hierarchical cluster analysis (HCA) in an analysis of the nutritional value, X-ray diffraction properties, texture, and sensory characteristics for normal ”rasgulla” samples and pineapple with different drying process. The authors observed approximately similar trends in the results obtained by the three compared algorithms.…”