Principal Component Analysis Visualizations in State Discovery by Animating Exploration Results

Sirola, Miki; Rinta‐Koski, Olli‐Pekka; Nguyen, Le Ngu; Hollmén, Jaakko

doi:10.1109/smartcomp55677.2022.00064

Cited by 2 publications

(4 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To clarify the state transitions here even more, in Figure 9, the statistical state transitions in the same analysis are shown, and, finally, in Figure 10, the probabilities of each state at each moment during the period of examination. In our prior publications [1,2], we highlighted animations as practical and effective tools for exploring and understanding states, thus providing physical interpretations of the process. Figure 8 showcases an animated structure using the pure flux dataset including only eight chosen variables from it and illustrating three distinct states, this time not clearly in chronological sequence.…”

Section: State Discovery Exploration and Visualization Resultsmentioning

confidence: 99%

“…In our prior publications [ 1 , 2 ], we highlighted animations as practical and effective tools for exploring and understanding states, thus providing physical interpretations of the process. Figure 8 showcases an animated structure using the pure flux dataset including only eight chosen variables from it and illustrating three distinct states, this time not clearly in chronological sequence.…”

Section: State Discovery Exploration and Visualization Resultsmentioning

confidence: 99%

“…However, in this study, we do not try to carry out any predictions, only classifications of state-based characteristics. Based on our earlier studies [ 1 ], we could use, for instance, the Hidden–Markov model [ 30 ] to predict states in future. The bookkeeping of farming field activities could be utilized as well.…”

Section: Discussionmentioning

confidence: 99%

“…The initial methodological concept used in this article is partly developed in our earlier work [ 1 ]. This study is an extension of the work [ 2 ], where pit soil data is analyzed, and process states are discussed.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Tracing State Structure for Ecological Processes in Soil Including Greenhouse Gas Exchange with Lower Atmosphere

Sirola,

Koskinen,

Polvinen

et al. 2024

Sensors

Self Cite

View full text Add to dashboard Cite

Exploring data aids in the comprehension of the dataset and the system’s essence. Various approaches exist for managing numerous sensors. This study perceives operational states to clarify the physical dynamics within a soil environment. Utilizing Principal Component Analysis (PCA) enables dimensionality reduction, offering an alternative perspective on the spring soil dataset. The K-means algorithm clusters data densities, forming the groundwork for an operational state description. Soil data, integral to an ecosystem, entails evident attributes. Employing dynamic visualization, including animations, constitutes a vital exploration angle. Greenhouse gas variables have been added to PCA to achieve more understanding in the interconnection of gas exchange and soil properties. Pit data and flux data are analysed both separately and together using a data-driven approach. The results look promising, showing the potential to add new values and more detailed state structures to ecological models. All experiments are conducted within the Jupyter programming environment, utilizing Python 3. The relevant literature on data visualization is examined. Through combined techniques and tools, the potential features of the soil ecosystem are observed and identified.

show abstract

Section: State Discovery Exploration and Visualization Resultsmentioning

confidence: 99%

Section: State Discovery Exploration and Visualization Resultsmentioning

confidence: 99%