Odor recognition series

One method of controlling response perseveration and other anticipation factors is to use a forced choice response indication based on two or more response categories. In the measurement of odors the panelist has to report the temporal position of positive stimuli in a series of random blanks. If the concentration is below the threshold, the test subjects will guess. As the odorant concentration will increase, the relative cumulative frequency for identification of the correct sample will be greater. In order to determine the relative odor recognition a correction must be made. 

The sensory evaluation differentiates between the stimulation threshold (a just detectable level where a perceptible but not yet definable deviation of the sample from the standard is observed) and the recognition threshold, a level where the odor is identifiable or creates odor problems (a no longer tolerable quality deterioration caused by a definite off odor and/or taste). The difference between a perceptible and identifiable level is usually only one to two steps of a geometric dilution series. Therefore, only undifferentiated odor and taste thresholds are given in Table 13-6, because of the very different sensitivities of individual testers. The perceptible (stimulation) levels of a less sensitive tester can overlap with the identifiable (recognition) level of another more sensitive tester. 

Threshold This is for the recognition of taste, odor, and flavor components. A series of solutions in order of physical concentration of the stimulus is used to determine the absolute threshold (ascending forced choice). 

Bicego used the similarity-based representation of electronic nose measurements for odor classification with the SVM method.In the similarity-based representation, the raw data from sensors are transformed into pairwise (dis)similarities, i.e., distances between objects in the dataset. The electronic nose is an array of eight carbon black-polymer detectors. The system was tested for the recognition of 2-propanol, acetone, and ethanol, with 34 experiments for each compound. Two series of 102 experiments were performed, the first one with data recorded after 10 minutes of exposure, whereas in the second group of experiments, the data were recorded after 1 second of exposure. The one-versus-one cross-validation accuracy of the first group of experiments was 99% for similarity computed using the Euclidean metric. For the second group of experiments, the accuracy was 79% for the Euclidean metric and 80% for the Manhattan metric. 

---