Table 6 2 Samples of the collected Turkish tweets.
Related Figures (16)
models are shown in Figure 4.1. so the nearby words are more weighted. The general architectures of the Word2Vec Figure 5.2. Structure of proposed CNN-based approach (testing and real-time) Figure 5.2 shows the mechanism of testing and/or using the approach in real- 6.1.2 Turkish Dataset Table 6. 1 Sample of the English dataset's Tweets (Figure Eight, 2015). Since we could not find any dataset in Turkish, we had to create the dataset Figure 6.2. The accuracy and F1 score of the studied algorithms using the third and fourth datasets. Table 6. 6 The execution time of each ensemble system for English. Experiment 5: Performance of Ensemble Systems for Turkish Language Table 6. 8 Execution time for ML algorithms for Turkish without Grid Search Table 6. 9 The accuracy of the studied Ensemble systems using Turkish datasets. 6.3.2 Investigating the Performance of Developed CNN Model Table 6. 10 The execution time of each ensemble system for Turkish datasets. In this section, the experiments for investigating the performance of tk developed CNN model and its results have been analyzed in detail. The main aims of this experiment are to find the degree that performance can Experiment 3: Investigating the Effect of the Input Window Size. The aim of this experiment is to investigate the effect of changing the number Table 6. 13 The effect of changing the number of words in the processed text window. Experiment 4: Effect of Changing the Number of CNN’s Filters Figure 6.3. The Accuracy, Precision, Recall, and Fl Score for the Developed System.
