Читать книгу Advanced Analytics and Deep Learning Models - Группа авторов - Страница 64

In this demonstration, they analyzed discrete arrangements. Those are mainly based on aspect-based sentiment analysis. The results we can see in Tables 3.2 and 3.3. They stated the results picked up with AFINN sentiment analysis algorithm, due to space reasons. It did not come out with any major dissimilarity with the CoreNLP algorithm. As it is based on CF user-based approach, on Yelp dataset and Tripadvisor dataset, they took top 10 aspects from the datasets. Besides, the above results are better than the previous 50 aspects. Accordingly, they did not take a bigger space.

One more attractive result comes from the Yelp and TripAdvisor by use of sub-aspect which gave a significant improvement in performance. Here, the maximum efficiency came by using the top 50 aspects. For a better understanding of this, we need to do further investigations [5, 20].

Table 3.2 Result comparison.

Result of MCRS-Based CF Experiment 1

Result of Experiment 2

Configuration	Dataset			Dataset
#neigh.	#asp.	Sub-asp	Yelp	TripAdvisor	Amazon		Configuration	Yelp	TripAdvisor	Amazon
10	10	Y	0.8362	0.7111	0.6464		Multi-U2U	0.8362	0.7111	0.6276
10	10	N	0.841	0.7564	0.6335		U2U-Euclidean	0.886	0.8337	0.7254
10	50	Y	0.841	0.7269	0.6346		U2U-Pearson	0.964	1.1222	0.9789
10	50	N	0.8364	0.8007	0.6276		Static-Multi-U2U	N.A.	0.798	N.A.
30	10	Y	0.8461	0.7677	0.712		Multi-I2I	0.864	0.8245	0.811
30	10	N	0.8473	0.7722	0.7122		I2I-Euclidean	0.8745	0.8429	0.8117
30	50	Y	0.8474	0.7743	0.7101		I2I-Pearson	1.1794	0.8644	0.9679
30	50	N	0.8494	0.8003	0.714		Static-Multi-I2I	N.A.	0.8474	N.A.
80	10	Y	0.8579	0.7971	0.7584		RatingSGD	0.8409	0.745	0.8859
80	10	N	0.8592	0.7953	0.7554		ParallelSGD	0.8409	0.7449	0.8852
80	50	Y	0.859	0.7907	0.7544		ALSWR	0.9545	0.9053	1.0354
80	50	N	0.8597	0.7995	0.7544

Result of MCRS Item–Based CF

Top 10 Main Aspects Extracted

Configuration	Dataset		Yelp	Place, food, service, restaurant, price menu, staff, drink, and lunch
#asp.	Sub-asp	Yelp	TripAdvisor	Amazon
10	Y	0.864	0.8245	0.811		TripAdvisor	Hotel, room, staff, location, service, breakfast, restaurant, bathroom, price, view
10	N	0.8643	0.8252	0.8117
50	Y	0.8641	0.8254	0.8118		Amazon	game, graphic, story, character, player price, gameplay, controller level, and music
50	N	0.8648	0.826	0.8124

It is noted that, to get better efficiency, we need to configure all the datasets in top 10 neighbors.

In the next step, they compared the algorithm with matrix factorization (MF) algorithms for getting better baselines. Besides, it is specified that TripAdvisor dataset has an unchangeable set of six features. For every analysis like “Cleanliness”, “location”, “value”, “service”, “sleep quality”, and “overall”. They have differentiated their way with a MCRS algorithm that depends on those aspects. In Table 3.5, we can see that their method gets the better of against all the baselines. These results surely established that their perception, since they proved that their approach could overpower both strategies exploiting single ratings and MCRS algorithms.

When multiple-criteria user-to-user CF is used as recommender algorithm, then the best overall results are obtained [5].