RESULTS AND DISCUSSION weights in performing calculations so as to obtain ⎪ 1 1 1 1 0.8 1 1 ⎪ 1 ⎪ the best alternative results using the Simple ⎪0.8 0.8 1 0.8 0.8 0.8 1 0.8⎪
RESULTS AND DISCUSSION. The electrooligomerization of enantiopure inherently chiral monomers on substrates like gold or glassy carbon in different media (aqueous media and organic solvents) leads to enanti- opure surfaces with a well-pronounced enantioselection abil- ity18-21. In order to confirm in a first step that the oligomers show chiral recognition properties also when they are generated on a Ppy film, we carried out tests with Ppy films, modified with (R) or (S) enantiopure oligomer layers, analog to what has been re- ported earlier25. The inherently chiral monomer chosen for this purpose was 2,2’-bis[2-(5,2’-bithienyl)]-3,3’-bithianaphthene (BT2T4), already well-known and characterized by Xxxxxxx et al. 18-21. It is constituted of 3,3’-bibenzothiophene as an atro- pisomeric central scaffold, with a sufficiently high racemization barrier to yield stable enantiomers, and functionalized with two 2,2’-bithiophenic units featuring two α-homotopic positions of the terminal thiophene rings, granting a regioregular oligomer- ization. In preliminary tests, Ppy films were deposited by chronopotentiometry on gold coated glass slides and then cov- ered with oligo-(R) or oligo-(S)-BT2T4 according to a literature procedure30-31. The final hybrid films were peeled off from the gold-coated glass slide and used for a first, more classic enanti- orecognition tests, performed by differential pulse voltammetry (DPV) in the presence of aqueous solutions of 5 mM L- or D- DOPA and lithium perchlorate 0.1 M. Figure 1. Top: Chemical structures of the enantiopure oligo- (S)- BT2T4 (A) and oligo-(R)-BT2T4 (B) molecules employed in the enantiorecognition tests. Bottom: Differential pulse voltam- metry signals of the enantioselective electrooxidation of 5 mM L- or D-DOPA in water and 0.1 M LiClO4 on a hybrid polymer layer composed of A) Ppy + oligo-(S)-BT2T4 and B) Ppy + ol- igo-(R)-BT2T4. From Figure 1A, it is obvious that the (S)-hybrid surface is preferentially reacting with L-DOPA, with a peak-to-peak sep- aration between the two enantiomers of about 200 mV. The re- sults are also perfectly coherent when the opposite configura- tion of the BT2T4 oligomer is deposited. The (R)-hybrid allows a preferential oxidation of D-DOPA. Consequently, at ~0.45 V vs. AgCl/Ag it is possible to selectively electrooxidize only one of the two DOPA enantiomers, depending on which oligomer configuration is deposited on Ppy. If this hybrid material is used as an electrode in a bipolar elec- trochemistry experiment, i...
RESULTS AND DISCUSSION. In this section, we report the results of the experimental phase conducted on XXXXX Xxxxxxx. The dataset comprises 8477 im- ages containing visual elements. In particular, the dataset in- cludes “embedded scenes”, i.e. images that in addition to the object of interest (e.g. a monument) contain other objects that di- rectly influence sentiment (e.g. happy people, waste, vandalism, and so on). The true sentiment is not automatically judged by the accompanying texts or hash-tags but has been manually estimated by human annotators, thus providing a more precise dataset. The experiments are based only on these images of the dataset, where both annotators have agreed on the sentiment. By remov- The application of our approach to this dataset yields good results in terms of precision, recall and F1-score and demonstrates the effectiveness of the proposed approach (Table 2). ing pictures with ambiguous sentiment, we increase the quality of the dataset and ensure the validity of the experiments. The dataset DCNNs Accuracy Precision Recall F1-score is split into 80% training and 20% test images. VGG16 0.65 0.67 0.65 0.66 ResNet 0.70 0.70 0.70 0.70 IncResNet 0.57 0.62 0.57 0.52 Figure 3. Confusion matrix for VGG16 network (learning rate=0.00001, batch=5, epochs=100). Table 2. Overall classification results comparing the three different DCNNs.
RESULTS AND DISCUSSION. In this section, the results of the experiments conducted on “Pho- tovoltaic images Dataset” are reported. This dataset presents un- balanced data, i.e. the classes are not represented equally. This is a scenario where the number of observations belonging to one class is significantly lower than those belonging to the other classes. Since classification algorithms tend to produce unsatisfactory clas- sifiers when faced with unbalanced datasets (Chawla, 2009), in precision = tp tp + fp • F1-score: is a measure of a test’s accuracy.
RESULTS AND DISCUSSION