研究目的
To determine the optimal nearest neighbor number, distance rule and classifier rule for PCA-EOC-KNN model for automated detection of NS1 fingerprint from SERS spectra of adulterated saliva.
研究成果
The PCA-EOC-KNN classifier performs with accuracy, precision, sensitivity, and specificity above 90% using Consensus classifier rule, Euclidean or Correlation or Cosine distance rule, and k-values of 1, 3, and 5. This model is effective for automated detection of NS1 from SERS spectra, benefiting early and non-invasive disease detection.
研究不足
The study does not specify limitations in the provided text, but potential areas for optimization could include the sensitivity to noise in kNN for small k-values and the dependence on the choice of distance and classifier rules.
1：Experimental Design and Method Selection:
The study uses a PCA-EOC-KNN classifier model for automated detection of NS1 from SERS spectra. PCA is used for feature extraction to reduce data dimensionality, and kNN is employed for classification with variations in k-value, distance rule, and classifier rule to find the optimal model.
2：Sample Selection and Data Sources:
Data are sourced from the UITM-NMRR-12-1278-12868-NS1-DENV database, consisting of 64 Raman spectra from control samples (without NS1) and 64 Raman spectra from adulterated samples (with NS1).
3：1). List of Experimental Equipment and Materials:
3. List of Experimental Equipment and Materials: Not specified in the provided text.
4：Experimental Procedures and Operational Workflow:
Spectra are pre-processed with background subtraction, baseline removal, smoothing, and normalization. PCA is applied for feature extraction. The dataset is split into 80% for training and 20% for testing. kNN classification is performed in MATLAB version 2018b, varying k-value (1 to 31), distance rules (Euclidean, Cityblock, Cosine, Correlation), and classifier rules (Nearest, Random, Consensus).
5：Data Analysis Methods:
Performance is evaluated based on accuracy, precision, sensitivity, and specificity metrics.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
