Table 1.
Comparison with the related studies.
Fig 1.
Framework for marketable value estimation of patents.
[a: The topic allocation probability used as an independent variable of the base-learner is calculated through LDA-based topic modeling. b: The market value of a patent used as a dependent variable is calculated through the event study methodology. c: Random Forest, MLP, and CNN are used as base learners. d: SVR is used as the algorithm of the ensemble model that calculates the final predicted value through the predicted value of the base-learner].
Fig 2.
Graphical model representation of the LDA model.
[a: K stands for the number of topics. b: M indicates the number of documents. c: N represents the total number of words in the document].
Table 2.
Model comparison.
Table 3.
Words and their weights for each topic.
Table 4.
Dominant topics for each document and the probability of being included in individual topics.
Fig 3.
Perplexity score results for selecting optimal topics.
[a: Perplexity score by topic. b: Perplexity score increase/decrease rate by topic].
Fig 4.
Coherence score results for selecting optimal topics.
[a: Coherence score by topic. b: Coherence score increase/decrease rate by topic].
Table 5.
Feature importance results using random forest.
Fig 5.
Regression results for the stock return and index return for the first eight patents.
[stock name compared to S&P 500 index—a: analog devices, b-c: applied materials, d: northrop grumman, e: life storage, f: tower semiconductor, g: Emerson Electric, h: GE].
Table 6.
Sample of abnormal return calculation results.
Table 7.
Basic statistics on the calculation results of abnormal returns for each subtechnical field of all patent document.
Table 8.
Hyperparameter tuning result by fold: Random forest.
Table 9.
Hyperparameter tuning result by fold: MLP.
Table 10.
Hyperparameter tuning result by fold: CNN.
Table 11.
Hyperparameter tuning result of meta learner (SVR).
Table 12.
Hyperparameter setting results of ensemble model.
Fig 6.
Changes in the loss rate according to an increase in epochs [a: prediction error of MLP model. b: prediction error of CNN model].
Table 13.
Hyperparameter tuning result of control group.
Fig 7.
Prediction comparison (horizontal axis: Prediction value, vertical axis: Target value).
[a: Linear regression model predicted value and target value scatterplot, b: Random forest model predicted value and target value scatterplot, c: MLP model predicted value and target value scatterplot, d: CNN model predicted value and target value scatterplot, e: Ensemble model predicted value and target value scatterplot].
Table 14.
Prediction error results for each model.
Table 15.
Results of normality test (Anderson-Darling normality test).
Table 16.
Results of the Wilcoxon rank-sum test with continuity correction.