The Impact of Simulated Spectral Noise on Random Forest and Oblique Random Forest Classification Performance

Agjee, Na’eem Hoosen; Mutanga, Onisimo; Peerbhay, Kabir; Ismail, Riyad

doi:10.1155/2018/8316918

Cited by 18 publications

(12 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The RF algorithm is commonly considered as largely resistant to noisy and mislabeled training data [3,9,27,28,42]. According to some studies, even features containing noise levels as high as 30% result in a decrease in Kappa accuracy in the range of 10%, which is considered a very moderate decrease [43]. Our results confirm the relative insensitivity of RF to noisy (errors) data.…”

Section: Impact Of the On-ground Reference Dataset Modification On Thsupporting

confidence: 79%

The t-SNE Algorithm as a Tool to Improve the Quality of Reference Data Used in Accurate Mapping of Heterogeneous Non-Forest Vegetation

Halladin-Dabrowska

Kania²,

Kopeć

2019

Remote Sensing

View full text Add to dashboard Cite

Supervised classification methods, used for many applications, including vegetation mapping require accurate "ground truth" to be effective. Nevertheless, it is common for the quality of this data to be poorly verified prior to it being used for the training and validation of classification models. The fact that noisy or erroneous parts of the reference dataset are not removed is usually explained by the relatively high resistance of some algorithms to errors. The objective of this study was to demonstrate the rationale for cleaning the reference dataset used for the classification of heterogeneous non-forest vegetation, and to present a workflow based on the t-distributed stochastic neighbor embedding (t-SNE) algorithm for the better integration of reference data with remote sensing data in order to improve outcomes. The proposed analysis is a new application of the t-SNE algorithm. The effectiveness of this workflow was tested by classifying three heterogeneous non-forest Natura 2000 habitats: Molinia meadows (Molinion caeruleae; code 6410), species-rich Nardus grassland (code 6230) and dry heaths (code 4030), employing two commonly used algorithms: random forest (RF) and AdaBoost (AB), which, according to the literature, differ in their resistance to errors in reference datasets. Polygons collected in the field (on-ground reference data) in 2016 and 2017, containing no intentional errors, were used as the on-ground reference dataset. The remote sensing data used in the classification were obtained in 2017 during the peak growing season by a HySpex sensor consisting of two imaging spectrometers covering spectral ranges of 0.4-0.9 µm (VNIR-1800) and 0.9-2.5 µm (SWIR-384). The on-ground reference dataset was gradually cleaned by verifying candidate polygons selected by visual interpretation of t-SNE plots. Around 40-50% of candidate polygons were ultimately found to contain errors. Altogether, 15% of reference polygons were removed. As a result, the quality of the final map, as assessed by the Kappa and F1 accuracy measures as well as by visual evaluation, was significantly improved. The global map accuracy increased by about 6% (in Kappa coefficient), relative to the baseline classification obtained using random removal of the same number of reference polygons.Data gathered based on the photointerpretation of a high-resolution orthophoto map are used less frequently [3]. Regardless of the method used to collect reference data, they may contain errors, even those acquired by authoritative sources [4]. Errors in the data may result from various factors, one of which being differences in the methods used for the identification of vegetation type when they come from different sources and when they were collected for different purposes [1]. In this case, the problem may arise from both a different interpretation of a given vegetation type by individual researchers, and the determination of vegetation units at different hierarchy levels (upper hierarchy level = higher level of generality). Errors in the reference data...

show abstract

Section: Impact Of the On-ground Reference Dataset Modification On Thsupporting

confidence: 79%

The t-SNE Algorithm as a Tool to Improve the Quality of Reference Data Used in Accurate Mapping of Heterogeneous Non-Forest Vegetation

Halladin-Dabrowska

Kania²,

Kopeć

2019

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…Random forest (RF) models were developed to predict when an animal was within 5 hours of calving using single variables and then combined variables. Random forest classifiers are ensemble machine learning algorithms which are considered to be more accurate than single classifiers, and more robust to noise (Agjee et al, 2018).…”

Section: Discussionmentioning

confidence: 99%

Using animal-mounted sensor technology and machine learning to predict time-to-calving in beef and dairy cows

Miller

Mitchell

Barker

et al. 2020

Animal

View full text Add to dashboard Cite

Worldwide, there is a trend towards increased herd sizes, and the animal-to-stockman ratio is increasing within the beef and dairy sectors; thus, the time available to monitoring individual animals is reducing. The behaviour of cows is known to change in the hours prior to parturition, for example, less time ruminating and eating and increased activity level and tail-raise events. These behaviours can be monitored non-invasively using animal-mounted sensors. Thus, behavioural traits are ideal variables for the prediction of calving. This study explored the potential of two sensor technologies for their capabilities in predicting when calf expulsion should be expected. Two trials were conducted at separate locations: (i) beef cows (n = 144) and (ii) dairy cows (n = 110). Two sensors were deployed on each cow: (1) Afimilk Silent Herdsman (SHM) collars monitoring time spent ruminating (RUM), eating (EAT) and the relative activity level (ACT) of the cow, and (2) tail-mounted Axivity accelerometers to detect tail-raise events (TAIL). The exact time the calf was expelled from the cow was determined by viewing closed-circuit television camera footage. Machine learning random forest algorithms were developed to predict when calf expulsion should be expected using single-sensor variables and by integrating multiple-sensor data-streams. The performance of the models was tested using the Matthew’s correlation coefficient (MCC), the area under the curve, and the sensitivity and specificity of predictions. The TAIL model was slightly better at predicting calving within a 5-h window for beef cows (MCC = 0.31) than for dairy cows (MCC = 0.29). The TAIL + RUM + EAT models were equally as good at predicting calving within a 5-h window for beef and dairy cows (MCC = 0.32 for both models). Combining data-streams from SHM and tail sensors did not substantially improve model performance over tail sensors alone; therefore, hour-by-hour algorithms for the prediction of time of calf expulsion were developed using tail sensor data. Optimal classification occurred at 2 h prior to calving for both beef (MCC = 0.29) and dairy cows (MCC = 0.25). This study showed that tail sensors alone are adequate for the prediction of parturition and that the optimal time for prediction is 2 h before expulsion of the calf.

show abstract

“…Five node splitting models (logistic regression, ridge regression, PLS, SVM and ensemble) of Oblique RF have performed significantly well. Oblique RF is a novel classification technique that is essentially an upgrade of the RF model (Agjee et al 2018). This model splits the feature space by means of various hyperplanes which are oblique in nature and can manage the noisy data more precisely (Do et al 2010;Menze et al 2011).…”

Section: Discussionmentioning

confidence: 99%

“…For that reason, it has better suited for the remote sensing data related spatial prediction. Besides, as this model uses supervised linear kind of models like ridge regression, SVM and others which are already stronger in multivariate node splitting at each node therefore the robustness and accuracy level of the Oblique RF model has gone better (Agjee et al 2018).…”

Section: Discussionmentioning

confidence: 99%

Head-cut gully erosion susceptibility modelling based on ensemble Random Forest with oblique decision trees in Fareghan watershed, Iran

Pham

Mukherjee²,

Norouzi

et al. 2020

Geomatics, Natural Hazards and Risk

View full text Add to dashboard Cite

The Impact of Simulated Spectral Noise on Random Forest and Oblique Random Forest Classification Performance

Cited by 18 publications

References 38 publications

The t-SNE Algorithm as a Tool to Improve the Quality of Reference Data Used in Accurate Mapping of Heterogeneous Non-Forest Vegetation

The t-SNE Algorithm as a Tool to Improve the Quality of Reference Data Used in Accurate Mapping of Heterogeneous Non-Forest Vegetation

Using animal-mounted sensor technology and machine learning to predict time-to-calving in beef and dairy cows

Head-cut gully erosion susceptibility modelling based on ensemble Random Forest with oblique decision trees in Fareghan watershed, Iran

Contact Info

Product

Resources

About