Correction: Park et al. Automatic Movie Tag Generation System for Improving the Recommendation System. Appl. Sci. 2022, 12, 10777

In the original publication [1], there was a mistake in Tables 2 and 4 and some texts related to the results of Tables 2 and 4. The error was caused by the incorrect selection of the dataset for testing, so we re-experimented by selecting a test dataset and adding a random seed value for reconstruction. The corrected

Table 2. Accuracy and learning time when running 100 epochs.

	Accuracy (Initial Epochs)	Max Accuracy	Learning Time
ResNet34	28.0822%	38.3562%	1311.7810 s
VGG-19	21.9178%	30.1370%	4423.7360 s
MobileNet	22.6027%	32.8767%	990.3590 s

appears below.

A correction has been made to Section 4.2, Paragraph 2, “933” should be replaced by “1013”.

A correction has been made to Section 5.1, Paragraph 1, “935” should be replaced by “972” and Paragraph 1, the phrase “30% of the total” was changed to “15% as validation data, and 15% of the total”.

In Section 5.1, Paragraph 2, result values were changed. The correct paragraph appears below:

Table 2 shows the accuracy and learning time of 100 epochs using ResNet34, VGG-19, and MobileNet-v2 models. MobileNet-v2 had 22.6027% accuracy at the first session, and the maximum accuracy is 32.8767%. VGG-19 had 21.9178% accuracy at the first session, and the maximum accuracy is 30.1370%, which is about 2.7% lower than MobileNet-V2. In the case of ResNet34, the first session of accuracy is 28.0822% and the maximum accuracy is 38.3562%, which was about 8.2% higher than VGG-19. As a basis for completing 100 epochs, MobileNet-v2 showed a learning speed of 990.3590 s, VGG-19 showed 4423.7360 s, and ResNet34 showed a learning speed of 1311.7810 s. ResNet34 showed the highest performance in accuracy and MobileNet-v2 showed the highest performance in speed in this study.

The corrected

Table 4. Comparison of real tags and tags generated from the proposed system (15 randomly selected in dataset). PM means partially matched; AM means all matched.

No.	Title		Genre	Movie Tag	Country	Result
1	Assassins	Real Tag	Action, Crime, Thriller	violence, romantic, suspenseful, sadist	US	PM
		Proposed Tag	Crime	violence, romantic	Anglosphere
2	Beautiful Girls	Real Tag	Comedy, Drama, Romances	violence, romantic, humor, atmospheric	US	PM
		Proposed Tag	Comedy	romantic	Anglosphere
3	Blood sport 3	Real Tag	Action, Sport	violence	US	PM
		Proposed Tag	Action	violence, comedy	Anglosphere
4	Bottle Rocket	Real Tag	Comedy, Crime, Drama	romantic, humor	US	PM
		Proposed Tag	Comedy	romantic, comedy	Anglosphere
5	Braveheart	Real Tag	Biography, Drama, History	violence, romantic, action, dramatic, inspiring	US	PM
		Proposed Tag	Drama	violence, suspenseful, atmospheric	Anglosphere
6	Desperado	Real Tag	Action, Crime, Thriller	violence, romantic, comedy, humor, action, mystery	US	PM
		Proposed Tag	Action	violence	Anglosphere
7	Drive	Real Tag	Action, Adventure, Comedy	violence, suspenseful	US	PM
		Proposed Tag	Action	violence, comedy	Anglosphere
8	Escape from L.A.	Real Tag	Action, Adventure, Sci-Fi	violence, humor, mystery	US	PM
		Proposed Tag	Action	violence, comedy	Anglosphere
9	Executive Decision	Real Tag	Action, Adventure, Thriller	violence, suspenseful, mystery	US	PM
		Proposed Tag	Action	violence	Anglosphere
10	House Arrest	Real Tag	Comedy, Family	romantic	US	AM
		Proposed Tag	Comedy	romantic	Anglosphere
11	In the Mouth of Madness	Real Tag	Drama, Fantasy, Horror	violence, comedy, suspenseful, fantasy	US	PM
		Proposed Tag	Drama	violence	Anglosphere
12	Kids	Real Tag	Drama	violence	US	PM
		Proposed Tag	Drama	violence, suspenseful	Anglosphere
13	Leaving Las Vegas	Real Tag	Drama, Romance	romantic, dramatic, dark, atmospheric, depressing	US	PM
		Proposed Tag	Drama	romantic	Anglosphere
14	Mallrats	Real Tag	Comedy, Romance	romantic, comedy, humor, comic	US	PM
		Proposed Tag	Comedy	violence, comedy	Anglosphere
15	Tommy Boy	Real Tag	Adventure, Comedy	comedy, humor	US	PM
		Proposed Tag	Comedy	comedy	Anglosphere

appears below.

A correction has been made to Section 5.4, Paragraph 1. The correct paragraph appears below:

The proposed automatic tag generation system analyzes sound and voice by inputting movie data. We selected 20 mood-related tags from about 70 tags and performed a multilabel classification on the 780 movies that were multi-labeled with those tags. The split of the dataset was carried out in the ratio of 70%, 15%, and 15% for the training, validation, and test datasets, respectively. Predicted tag is generated based on the analysis, and the tag list is stored as a CSV file. As shown in Table 4, only 15 lists were randomly extracted and summarized as results. Table 4 summarizes the matching results of the actual tags provided in the dataset and the automatically generated tags proposed in this paper.

A correction has been made to Section 5.4, Paragraph 2. The correct paragraph appears below:

Our work also shows an accuracy score result value of 86.84% for the test dataset. Although there is a difference in the number of actual tags, it can be confirmed that the extracted tags are included in the actual tags. The number of tags can be solved by increasing classification items or adding keywords in the system. Video platforms are already used by many consumers worldwide, and platform companies are continuously studying recommendation systems to provide customized services to consumers. Automating generation of tags for movie content through the methods proposed in this paper will help platform companies that provide recommendation services to reduce the time and cost for matching content to consumers.

The authors state that the scientific conclusions are unaffected. This correction was approved by the Academic Editor. The original publication has also been updated.