Evaluation of Non-linearity in MIR Spectroscopic Data for Compressed Learning

Mid-Infrared (MIR) spectroscopy has emerged as the most economically viable technology to determine milk values as well as to identify a set of animal phenotypes related to health, feeding, well-being and environment. However, Fourier transform-MIR spectra incurs a significant amount of redundant data. This creates critical issues such as increased learning complexity while performing Fog and Cloud based data analytics in smart farming. These issues can be resolved through data compression using unsupervisory techniques like PCA, and perform analytics in the compressed-domain i.e. without decompressing. Compression algorithms should preserve non-linearity of MIRS data (if exists), since emerging advanced learning algorithms can improve their prediction accuracy. This study has investigated the non-linearity between the feature variables in the measurement-domain as well as in two compressed domains using standard Linear PCA and Kernel PCA. Also, the non-linearity between the feature variables and the commonly used target milk quality parameters (Protein, Lactose, Fat) has been analyzed. The study evaluates the prediction accuracy using PLS and LS-SVM respectively as linear and nonlinear predictive models.

[1]  Weilin Huang,et al.  Linear and nonlinear dimensionality reduction for face recognition , 2009, 2009 16th IEEE International Conference on Image Processing (ICIP).

[2]  Yande Liu,et al.  On-line NIR predicting soluble solids content of intact pears combination with wavelet transform and support vector regression , 2010, 2010 Sixth International Conference on Natural Computation.

[3]  Chamil Kulatunga,et al.  Opportunistic Wireless Networking for Smart Dairy Farming , 2017, IT Professional.

[4]  L. J. P. van der Maaten,et al.  An Introduction to Dimensionality Reduction Using Matlab , 2007 .

[5]  J. V. Anand,et al.  Regression based analysis of effective hydrocast in underwater environment , 2014, TENCON 2014 - 2014 IEEE Region 10 Conference.

[6]  A. Brodkorb,et al.  Prediction of bovine milk technological traits from mid-infrared spectroscopy analysis in dairy cows. , 2015, Journal of dairy science.

[7]  Y. Chen,et al.  Combination of the Manifold Dimensionality Reduction Methods with Least Squares Support vector machines for Classifying the Species of Sorghum Seeds , 2016, Scientific Reports.

[8]  Chamil Kulatunga,et al.  Fog-enabled WSN system for animal behavior analysis in precision dairy , 2017, 2017 International Conference on Computing, Networking and Communications (ICNC).

[9]  Klemens Böhm,et al.  Multivariate Maximal Correlation Analysis , 2014, ICML.

[10]  Min Chen,et al.  On the computation offloading at ad hoc cloudlet: architecture and service modes , 2015, IEEE Communications Magazine.

[11]  K. Johana,et al.  Benchmarking Least Squares Support Vector Machine Classifiers , 2022 .

[12]  S. Wolfert,et al.  Big Data in Smart Farming – A review , 2017 .

[13]  Pascal Vincent,et al.  Representation Learning: A Review and New Perspectives , 2012, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[14]  Tansu Alpcan,et al.  Fog Computing May Help to Save Energy in Cloud Computing , 2016, IEEE Journal on Selected Areas in Communications.

[15]  P. Garthwaite.  An Interpretation of Partial Least Squares , 1994 .

[16]  Lijuan Cao,et al.  A comparison of PCA, KPCA and ICA for dimensionality reduction in support vector machine , 2003, Neurocomputing.

[17]  H. Vincent Poor,et al.  Dimensionally distributed learning models and algorithm , 2008, 2008 11th International Conference on Information Fusion.

[18]  Simon J. Sheather,et al.  Exploring data sets using partial residual plots based on robust fits , 1997 .

[19]  R. Calderbank,et al.  Compressed Learning : Universal Sparse Dimensionality Reduction and Learning in the Measurement Domain , 2009 .

[20]  Fabio Del Frate,et al.  Pixel Unmixing in Hyperspectral Data by Means of Neural Networks , 2011, IEEE Transactions on Geoscience and Remote Sensing.