Enhancing Crop Yield Prediction Utilizing Machine Learning on Satellite-Based Vegetation Health Indices.

Accurate crop yield forecasting is essential in the food industry's decision-making process, where vegetation condition index (VCI) and thermal condition index (TCI) coupled with machine learning (ML) algorithms play crucial roles. The drawback, however, is that a one-fits-all prediction model is often employed over an entire region without considering subregional VCI and TCI's spatial variability resulting from environmental and climatic factors. Furthermore, when using nonlinear ML, redundant VCI/TCI data present additional challenges that adversely affect the models' output. This study proposes a framework that (i) employs higher-order spatial independent component analysis (sICA), and (ii), exploits a combination of the principal component analysis (PCA) and ML (i.e., PCA-ML combination) to deal with the two challenges in order to enhance crop yield prediction accuracy. The proposed framework consolidates common VCI/TCI spatial variability into their respective subregions, using Vietnam as an example. Compared to the one-fits-all approach, subregional rice yield forecasting models over Vietnam improved by an average level of 20% up to 60%. PCA-ML combination outperformed ML-only by an average of 18.5% up to 45%. The framework generates rice yield predictions 1 to 2 months ahead of the harvest with an average of 5% error, displaying its reliability.

The Stressor in Adolescence of Menstruation: Coping Strategies, Emotional Stress & Impacts on School Absences among Young Women in Nepal.

Throughout the developing world, girls face hardships surrounding menstruation, often resulting in poor emotional wellbeing and missing school. Providing ways to keep girls in school will increase their educational and earning potentials, which will ultimately trickle down to improving the economic standing of nations in the next generation. Informed by the Transactional Model of Stress and Coping, this work evaluates the roles that cultural and school environments play in appraisals of menstruation as a major life stressor for adolescent females and the impacts of emotional stress on missing school. Using primary survey data from schools in Nepal, robust results are found to support the theoretical framework based on conditional mixed-process (CMP) estimation with fixed effects, utilizing multiple index building techniques. Strong cultural norms during menstruation appear to increase the probability of girls self-reporting emotional stress, while the presence of hygiene supporting infrastructure at schools reduces this outcome. Furthermore, there is strong support for the finding that the presence of emotional stress during menstruation increases the likelihood of not only missing school but also for an extended period of time. Our findings motivate increasing government policies to provide stronger hygiene infrastructure in schools to improve successful coping skills and attendance rates.

Comparison between synthetic and rosemary-based antioxidants for the deep frying of French fries in refined soybean oils evaluated by chemical and non-destructive rapid methods.

Using natural antioxidants instead of synthetics ones has been the tendency for retarding the oil deterioration during repeated deep frying process. Concerning this, the comparison between synthetic tertiarybutyl hydroquinone (TBHQ) and rosemary-based antioxidants in frying French fries was hereby evaluated. The quality and stability of frying oils with rosemary-based antioxidants showed higher efficiency than TBHQ regarding oxidation parameters (i.e., chemical indices, sensory, etc.), where rosmarinic acid (RA) was the most effective, followed by rosemary extracts (RE) and carnosic acid (CA). LF-NMR results were highly correlated (R2 = 0.909-0.998) to the change in physicochemical properties tested, where RA could effectively regulate the relaxation spectrum (T2) change and decrease single component relaxation time (T2W). The PCA graph of NIRS also revealed the dynamic change of antioxidant effectiveness in accordance with that obtained by chemical methods. Hence, both LF-NMR and NIRS can be expected as rapid and efficient methods for future monitoring the frying process.

Rapid fingerprinting technology of heavy oil spill by mid-infrared spectroscopy.

With the increase of unconventional oil production and transportation, the detection methods of light crude oil have been challenged. Mid-Infrared spectroscopy can reflect the functional group of the oil related samples, which has strong absorption signals with distinguishable peaks featured as a fast, economy, and robust technique. Nevertheless, the previous study and application of oil relevant samples, such as petroleum chemical industry online monitoring, are mainly based on Near-infrared spectroscopy. Recently, the rapid development of the spectral instrument manufacturing and the data analysis methods provides a more comprehensive technical support for the rapid and accurate identification of marine oil spill by Mid-infrared spectroscopy. In this paper, 10 crude oil samples were selected for infrared spectroscopy detection, and the results were analysed and compared with those of gas chromatography flame ionization detection method. The character information of the IR spectra and GC/FID chromatograms were extracted and classified both by principal component analysis and partial least squares regression. Under the condition of small sample size, the recognition accuracy was up to 100%. The results show that the mid-infrared method combined with chemometrics can be expected to achieve rapid, accurate and economical identification of heavy oil species.

Investigating sensory properties of seven watermelon varieties and factors impacting refreshing perception using quantitative descriptive analysis.

Watermelon (Citrullus lanatus) is known for its refreshing quality, though its sensory attributes have never been related to its perceived refreshment. Modified quantitative descriptive analysis by a trained panel was used to examine the sensory profile of seven watermelon varieties. Eleven attributes including perceived refreshing intensity were measured on a 0-10 line scale using chemical references. Watermelon samples were evaluated with and without nose clips to control orthonasal and retronasal aroma and temperature was included as a variable to observe their effects on perceived refreshment. The dominant watermelon attributes were wateriness, refreshing, crispness, sweet, mealiness, fresh, ripe, and melon. The varieties were best differentiated by refreshing (p < 0.001), crispness (p = 0.002), sweet (p < 0.001), mealiness (p = 0.016), green (p = 0.007), and sour perception (p < 0.001). Captivation and Excursion were the most refreshing varieties. Captivation, Excursion, and Seedless varieties were less refreshing when flavor perception was inhibited; ratings ranged from 6.8 to 7.2 without nose clips and 5.9-6.0 with nose clips (p = 0.002). Refreshing was most positively driven by wateriness, followed by crispness, fresh, melon, and sweet, and negatively driven by mealiness, as indicated by partial least square regression. Samples served cold were more refreshing (ratings of 7.1 without and 6.0 with nose clips) than those served at room temperature (ratings of 4.9 without and 3.5 with nose clips), p < 0.001. This study defined the sensory profile of seven watermelon varieties and showed that flavor, texture, and temperature were responsible for the refreshing perception of watermelon for the first time.

Temporal scatterplots.

Visualizing high-dimensional data on a 2D canvas is generally challenging. It becomes significantly more difficult when multiple time-steps are to be presented, as the visual clutter quickly increases. Moreover, the challenge to perceive the significant temporal evolution is even greater. In this paper, we present a method to plot temporal high-dimensional data in a static scatterplot; it uses the established PCA technique to project data from multiple time-steps. The key idea is to extend each individual displacement prior to applying PCA, so as to skew the projection process, and to set a projection plane that balances the directions of temporal change and spatial variance. We present numerous examples and various visual cues to highlight the data trajectories, and demonstrate the effectiveness of the method for visualizing temporal data.

Rapid and sensitive identification of uropathogenic Escherichia coli using a surface-enhanced-Raman-scattering-based biochip.

Rapid, selective and sensitive sensing of bacteria remains challenging. We report on a highly sensitive and reproducible surface-enhanced Raman spectroscopy (SERS)-based sensing approach for the detection of uropathogenic Escherichia coli (E. coli) bacteria in urine. The assay is based on the specific capture of the bacteria followed by interaction with cetyltrimethylammonium bromide (CTAB)-stabilised gold nanorods (Au NRS) as SERS markers. High sensitivity up to 10 CFU mL-1 is achieved by optimizing the capture interface based on hydrogenated amorphous silicon a-Si:H thin films. The integration of CH3O-PEG750 onto a-Si:H gives the sensing interface an efficient anti-fouling character, while covalent linkage of antibodies directed against the major type-1 fimbrial pilin FimA of the human pathogen E. coli results in the specific trapping of fimbriated E. coli onto the SERS substrate and their spectral fingerprint identification.

Path Loss Prediction based on Machine Learning Techniques: Principal Component Analysis, Artificial Neural Network and Gaussian Process.

Although various linear log-distance path loss models have been developed for wireless sensor networks, advanced models are required to more accurately and flexibly represent the path loss for complex environments. This paper proposes a machine learning framework for modeling path loss using a combination of three key techniques: artificial neural network (ANN)-based multi-dimensional regression, Gaussian process-based variance analysis, and principle component analysis (PCA)-aided feature selection. In general, the measured path loss dataset comprises multiple features such as distance, antenna height, etc. First, PCA is adopted to reduce the number of features of the dataset and simplify the learning model accordingly. ANN then learns the path loss structure from the dataset with reduced dimension, and Gaussian process learns the shadowing effect. Path loss data measured in a suburban area in Korea are employed. We observe that the proposed combined path loss and shadowing model is more accurate and flexible compared to the conventional linear path loss plus log-normal shadowing model.

Characterization of industrial plumes of volatile organic compounds by guided sampling.

Instead of manual sampling in a random way near a source area, this study used trigger sampling guided by an analyzer at a fixed site near a refinery plant to obtain the chemical composition of volatile organic compounds (VOCs) representative of the source. The analyzer was built in-house to measure total VOC (TVOC) levels by subtracting methane from total combustible organic compounds (TOC) with flame ionization detection. The analyzer with minute resolution provided instantaneous measurements of TVOCs to trigger canister sampling at the moments of VOC plumes in a source area. The chemical composition of the 13 trigger samples were compared with the other non-trigger samples randomly collected either within the refinery or on an urban street. All samples were analyzed by gas chromatography-mass spectrometry/flame ionization detection (GC-MS/FID) for detailed speciation. High agreement in total VOC abundance between the analyzer and GC-MS/FID indicates internal consistency of the two techniques and the robustness of the TVOC analyzer to guide sampling of VOC plumes. The trigger samples showing very high consistency in the overall composition imply that sampling at the right moments of plume arrivals can facilitate characterization of the source profiles, which can hardly be achieved by random sampling. The coupling of the fast-and-slow analytical techniques to guide sampling is proven to be an effective means to probe source characteristics.

[Comparison of the Geochemical Characteristics of Karst Springs of a Vertically Zoned Climate Region under Human Activity: A Case of Shuifang Spring and Bitan Spring in the Jinfo Mountain Area, Chongqing].

To investigate the hydrochemical variation of karstic groundwaters in a vertically zoned climate region affected by human activity, Shuifang Spring and Bitan Spring in the Jinfo Mountain area of Chongqing were selected as a study site. Based on the differences between the natural state and intensity of human activity of these two springs, their hydrogeochemical characteristics and the controlling factors on karstic groundwaters were analyzed by means of independent sample t tests, the Gibbs graphic method, principle component analysis (PCA), and geochemical susceptivity analysis. The results show that differences in karst development in the vertical climatic zone leads to higher total ion concentrations in Bitan Spring than in Shuifang Spring. The hydrochemical types of Shuifang Spring and Bitan Spring are HCO3-Ca and HCO3-Ca·Mg, respectively, which reflect the lithology of their different elevations. Carbonate rock dissolution is the main source of Ca2+, Mg2+, and HCO3- in karstic groundwaters. Hotel sewage discharge supplies SO42-, NO3-, PO43-, K+, and Na+ in Shuifang Spring, which peaked in winter and summer, while hydrochemical parameters of Bitan Spring changed smoothly throughout the year. The water quality of Bitan Spring is better than Shuifang Spring (Shuifang Spring water is classified as Class Ⅳ). PCA shows that the water-rock interaction was the first controlling factor. Hotel sewage discharge and ions from precipitation had important effects on Shuifang Spring and Bitan Spring, respectively. In addition, the effects of soil erosion and leaching caused by precipitation also impact on the water quality of two springs to some extent. The geochemical susceptibility of Shuifang Spring was greater than that of Bitan Spring; therefore, corresponding measures should be formulated according to the characteristics of these differently elevated karst systems when exploiting groundwater resources. This is especially the case for the treatment of hotel sewage.

Mass spectrometry-based metabolomics approach to reveal differential compounds in pufferfish soups: Flavor, nutrition, and safety.

Some Eastern Asian countries deem pufferfish - especially its muscle - a culinary delight. Herein, molecular mass fingerprinting of soups prepared by Takifugu flavidus, Takifugu obscurus and Takifugu rubripes was established via matrix-assisted laser desorption/ionization - mass spectrometry (MALDI MS). Soup samples were directly analyzed by MALDI MS to collect mass spectra within 0-700 Da in a quick way, followed by principal component analysis to distinguish the different soups and to find out the distinctive compounds among the soups. High performance liquid chromatography - tandem MS (HPLC-MS/MS) was applied to identify the compounds. Nineteen compounds were identified from the HPLC-MS/MS data by using METLIN database. Through literature mining, we found that these compounds are closely related to the flavor, nutrition, and safety of pufferfish soups. This method can also be used as a facile way to distinguish between different pufferfish fillets when morphological characters have been damaged or destroyed.

Street dust heavy metal pollution implication on human health in Nicosia, North Cyprus.

The consequence of carcinogenic and non-carcinogenic health risks of the heavy metal concentrations in street dust of North Cyprus is yet to be reported. This study is aimed at investigating the concentration of six different heavy metals' concentration explicitly: As, Cu, Zn, Ni, Cr, and Pb, along leading highways in Nicosia. The result obtained was analyzed using an X-ray fluorescent machine. Multivariate and statistical methods were applied for the data analysis using xlstat MS-excel; furthermore, index of geo-accumulation (Igeo) and human health risk assessment using exposure pathways as defined by United State Environmental Protection Agency (USEPA) pollution mode were also used for level assessment and health risk implications. The average (M ± SD) concentrations of the metals in the dust are as follows: As (17.48 ± 1.53 mg/kg), Cu (51.86 ± 8.60 mg/kg), Cr (321.14 ± 8.20 mg/kg), Pb (35.62 ± 8.54 mg/kg), Ni (64.79 ± 8.72 mg/kg), and Zn (136.13 ± 30.85 mg/kg). Variation coefficient, Vc, and principle component analysis (PCA) suggested that As, Cr, Ni, and Pb have same source of pollution emission from both natural and anthropogenic activities, Zn from traffic emission while Cu from natural source. However, the result was compared with other nearby towns bordering North Cyprus; all the metal shows similar pattern of pollution with the exception of Cr which is 5 and 11 times higher than street dust of Amman (Jordan) and Tokat (Turkey), respectively. Additionally, Igeo result has the following decreasing order: Zn > Cr > Pb > Ni > Cu > As and also revealed that the As, Cu, and Ni have originated from natural source. Cr has mix source: one from traffic and the other one from atmospheric deposition. Also, Pb is emitted from industrial pollution, whereas 80% of Zn are from traffic-related emissions. The non-carcinogenic health risk (hazard quotient (HQ) and hazard index (HI)) follows the order Cr > As>Ni > Pb > Zn > Cu for children and adults. It is found that the HI of As, Cu, Ni, Pb, and Zn is less one; hence, the street dust does not exhibit non-carcinogenic health risk. But that of Cr content is greater than one, with HI values of Cr 1.44E+02 and 1.55E+01 for children and adults, respectively. The result for carcinogenic health risk (total cancer risk (TCR)) has the following order: Pb (1.42E-05) > Cr (4.81E-09) > (Ni 1.35E-09) > As (1.96E-10). With all the values less than threshed hole limit of TCR ≥ 10-4, street dust does not possess carcinogenic health risk for the entire values of six heavy metals considered in this work.

Effect of crystallinity on pretreatment and enzymatic hydrolysis of lignocellulosic biomass based on multivariate analysis.

In this work, multivariate data analysis was employed to correlate variables of pretreatment process of lignocellulosic biomass. Principal component analysis and partial least square methods were performed to get the inner-relationship and data interpretation between the crystallinity and other parameters of mechanical refining-assisted sodium hydroxide pretreatment followed by enzymatic saccharification of corn stover. The PCA and PLS models showed that Sodium hydroxide dosage, mechanical refining treatment, lignin removal rate and crystallinity had close inner-related relationship with the efficiency of pretreatment and enzymolysis. Alkaline reaction and mechanical refining treatment had strong influence on the crystallinity. Multivariate data analysis revealed that pretreated corn stover samples with lower crystallinity were more easily hydrolyzed by enzyme and could get more final reducing sugar. This work could offer a new methodology to get further understanding of effect of crystallinity on the crop residue lignocellulosic biomass conversion process.

Spatially resolved analysis of liquid crystal displays by laser ablation-atmospheric pressure chemical ionization-mass spectrometry.

In recent years, liquid crystal displays (LCDs) became more and more popular in industrial and consumer electronics applications. Unevenly distributed liquid crystals (LC) and the presence of organic contaminations in the LC layer, for example from the display sealant, can cause display defects and therefore lower the display quality. This work describes the development of a fast and easy screening method for display samples by laser ablation coupled to atmospheric pressure chemical ionization-mass spectrometry (LA-APCI-MS). With this ambient technique, LC components as well as sealant materials, which are one of the main sources for display contaminations, can be detected simultaneously using spatial resolutions down to 100 µm. Minimal sample preparation is needed for this method and exact masses can be acquired by using an Orbitrap mass analyzer. By LA-APCI-MS, nineteen different LCs and five different sealant materials were detected as pure substances as well as on commercially available LCDs. In addition, a principle component analysis (PCA) was used to differentiate between LCs and sealant ions in the complex data matrix.

Variation in Ceratonia siliqua pod metabolome in context of its different geographical origin, ripening stage and roasting process.

Carob is a legume tree of a considerable commercial importance for the flavor and sweet industry. In this context, it is cultivated mostly for its pods, which are known for their nutritive value and multiple health benefits. However, metabolite patterns, underlying these properties are still mostly uncharacterized. In this study, the role of geographical origin, ontogenetic changes and thermal processing on the Ceratonia siliqua pod metabolome was assessed by mass spectrometry (MS)-based metabolomics. Thereby, a total of 70 fruits primary metabolites, represented mainly by carbohydrates, organic and amino acids were detected. Analysis of secondary bioactive metabolites assessed by ultra-high-performance liquid chromatography-electrospray ionization high resolution mass spectrometry (UHPLC-ESI-HR-MS) revealed in total 83 signals. The major signals, most significantly contributing in discrimination of C. siliqua specimens were assigned to tannins and flavonoids. PCA models derived from either UHPLC-MS or GC-MS proved to be powerful tools for discrimination of C. siliqua specimens.

Optimization of rendering process of poultry by-products with batch cooker model monitored by electronic nose.

The best available controlled technology for transforming the disposal of animal by-products and mortalities is rendering. Two aspects of rendering process are mentioned in this research; product quality and emissions. A model of batch cooker with temperature, pressure and agitator speed controllers was designed and developed in order to optimize the process and to investigate the effect of changes in rendering conditions on quality of poultry by-product meal and also on pollutant emissions. An electronic nose system was designed and built based on metal oxide semiconductor sensors to monitor the gases emitted from batch cooker model. Also, GC-MS was used to identify the emitted components. In order to optimize the rendering process, response surface methodology was performed on temperature, cooking time and agitator speed variables. Results showed that the temperature of 140 °C (internal pressure equivalent to about 3.2 bar), the cooking time of 45 min and the agitator speed of 20 rpm optimized the process of batch cooking to maximize the percentage of protein and minimize the percentage of fat, moisture content, energy consumption and emission of pollutants. By GC-MS analysis, about 100 compounds include hydrocarbons, volatile fatty acids, sulfur-containing compounds, alcohols, ketones, aldehydes, and furans were observed in the emission of a batch cooker model. The major groups were organic acids and amides. Principle component analysis showed the most suitable sensors for detecting unpleasant odors from rendering plants.

[Analysis of volatile and semivolatile components of burley tobacco before and after baking by headspace solid-phase microextraction gas chromatography/mass spectrometry with chemometrics].

The volatile and semivolatile components in burley tobacco leaves were analyzed by headspace solid-phase microextraction gas chromatography/mass spectrometry (HS-SPME-GC/MS). Twenty milligrams of tobacco powder was incubated at 60 ℃ for 8 min and then extracted using 65 µm polydimethylsiloxane/divinylbenzene (PDMS/DVB) fiber for 40 min. Finally, the fiber was desorbed at 250 ℃ for 3 min. One hundred twenty two kinds of volatile and semivolatile components in the burley tobacco leaves were tentatively identified by comparing with standard products and mass spectrometry databases, and these compounds were semi-quantitatively analyzed by the internal standard method. The differences between the volatile and semivolatile components for the burley tobacco leaves before and after baking were discriminated and visualized by principal component analysis (PCA) and partial least-squares discriminant analysis (PLS-DA). This method has the advantages of small sample size, simple pretreatment, and high sensitivity. In combination with chemometrics, the HS-SPME-GC/MS method is suitable for discriminating changes in the chemical composition of burley tobacco before and after baking, and hence has broad prospective application in the optimization of the baking conditions used for burley tobacco.

Novel Arsenic Markers for Discriminating Wild and Cultivated Cordyceps.

Ophiocordyceps sinensis has been utilized in China and adjacent countries for thousands of years as a rare functional food to promote health and treat diverse chronic diseases. In recent years, adulterants are usually identified in the processed products of wild O. sinensis. However, the effective adulteration examination has to be additionally performed except their routine test, and accordingly is time- and money-consuming. Recently, arsenic determination has become a necessary test for confirming whether the concentrations of inorganic arsenic are over the O. sinensis limit. In this work, the contents of total arsenic and As species in cultivated O. sinensis, Cordyceps militaris, and other edible fungi were determined by ICP-MS and HPLC-ICP-MS. The results suggest that the As speciation exhibits a species-specific behavior, and accompanies the effect of the As background. The proportions of unknown organic As and contents of total As may be considered as sensitive markers for discriminating wild O. sinensis. This result provides a novel clue for discriminating wild and artificially cultivated mushrooms/their products, with emphasis on arsenic markers for authenticating wild O. sinensis.

Helical Multi-walled Carbon Nanotubes as an Efficient Material for the Dispersive Solid-Phase Extraction of Low and High Molecular Weight Polycyclic Aromatic Hydrocarbons from Water Samples: Theoretical Study.

The differences in effectiveness of multi-walled carbon nanotubes (MWCNTs) as the dispersive solid-phase extraction (dSPE) sorbent for the selective extraction of polycyclic aromatic hydrocarbons (PAHs) were explained on the basis of theoretical study. It was observed that for low molecular weight PAHs, the recoveries using non-helical and helical MWCNTs were similar. In contrary, for PAHs containing five or more aromatic rings, the extraction efficiency was higher using HMWCNTs than for non-helical ones. Principle component analysis (PCA) as well as providing structural parameters and interaction energies for adsorption processes (PAH + CNT → PAH-CNT) have been used for this purpose. All the PAH + CNT → PAH-CNT adsorption processes considered were found to be thermodynamically favorable. However, the adsorption energies (Eads) for PAHs and the helical carbon nanotube surface estimated for the B(a)P-HCNT and I(1,2,3-cd)P-HCNT are substantially less negative than those observed for PAH molecules interacting with the non-helical CNT. Namely, the Eads calculated in simulated aqueous environment for the B(a)P-MWCNT(6,2) and I(1,2,3-cd)P-MWCNT(6,2) were respectively - 43.32 and - 59.98 kcal/mol, while values of only - 7.75 kcal/mol (B(a)P-HCNT) and - 9.13 kcal/mol (I(1,2,3-cd)P-HCNT) were found for the corresponding PAH-HCNT systems. Therefore, we conclude that the replacement of MWCNTs with HCNTs leads to PAH-HCNT systems in which the interaction energies are much smaller than those estimated for the corresponding PAH-MWCNT systems. HMWCNTs are therefore recommended as the dSPE sorbent phase for the extraction of both low and high molecular weight PAHs from water samples.

Analysis of Floral Volatile Components and Antioxidant Activity of Different Varieties of Chrysanthemum morifolium.

This study investigated the volatile flavor compounds and antioxidant properties of the essential oil of chrysanthemums that was extracted from the fresh flowers of 10 taxa of Chrysanthemum morifolium from three species; namely Dendranthema morifolium (Ramat.) Yellow, Dendranthema morifolium (Ramat.) Red, Dendranthema morifolium (Ramat.) Pink, Dendranthema morifolium (Ramat.) White, Pericallis hybrid Blue, Pericallis hybrid Pink, Pericallis hybrid Purple, Bellis perennis Pink, Bellis perennis Yellow, and Bellis perennis White. The antioxidant capacity of the essential oil was assayed by spectrophotometric analysis. The volatile flavor compounds from the fresh flowers were collected using dynamic headspace collection, analyzed using auto thermal desorber-gas chromatography/mass spectrometry, and identified with quantification using the external standard method. The antioxidant activities of Chrysanthemum morifolium were evaluated by DPPH and FRAP assays, and the results showed that the antioxidant activity of each sample was not the same. The different varieties of fresh Chrysanthemum morifolium flowers were distinguished and classified by fingerprint similarity evaluation, principle component analysis (PCA), and cluster analysis. The results showed that the floral volatile component profiles were significantly different among the different Chrysanthemum morifolium varieties. A total of 36 volatile flavor compounds were identified with eight functional groups: hydrocarbons, terpenoids, aromatic compounds, alcohols, ketones, ethers, aldehydes, and esters. Moreover, the variability among Chrysanthemum morifolium in basis to the data, and the first three principal components (PC1, PC2, and PC3) accounted for 96.509% of the total variance (55.802%, 30.599%, and 10.108%, respectively). PCA indicated that there were marked differences among Chrysanthemum morifolium varieties. The cluster analysis confirmed the results of the PCA analysis. In conclusion, the results of this study provide a basis for breeding Chrysanthemum cultivars with desirable floral scents, and they further support the view that some plants are promising sources of natural antioxidants.