ايگين كريم زاده

Digital innovation is the use of digital technology to solve business problems such as process optimization, improve customer experiences, products and present new business models. This study deals with how products develop products based on online customer satisfaction analysis using artificial intelligence algorithms, which is one of the digital innovation tools. In fact, this research examines the following research topics in the field of new product development: (1) How to use online data created by customers in digital space (websites, social media, etc.) And get customer demands for products? (2) How to analyze online data, preferences, and customer behavior using digital innovation tools such as artificial intelligence (genetic algorithm)? To explore these research topics, this research offers a new way to analyze customer data set for products and find the best features and importance based on customer satisfaction. This method initially uses an evolutionary algorithm (genetic algorithm) to find the most important and best combination of features that match the final rating of customers for the goods. Then, using machine learning algorithms, eva‎luates the solutions proposed by the genetic algorithm. Lastly, the Shap model, which is a way to explain the output of any machine learning algorithm, is used to interpret the model and identify the importance and value of each product feature in customer satisfaction. Using the suggested method, the best and most important compounds of product features that affect customer satisfaction were determined. Also, a table of values values for each product feature and different modes of a feature was identified. The relationship between product features and their impact on each other is also displayed. In general, quantitative and qualitative results, such as the optimal combination of features, performance score of each option, and explaining how each feature affects customer satisfaction. Finally, the results of the proposed method are compared with similar research and optimization is determined. Interpretation of the results of this study shows that the proposed method shows better performance in eva‎luation criteria than other designated models. It also covers uncertainty and incomplete information such as ranking and rating of online products.

نوآوري , ژنتيك الگوريتم , نوآوري ديجيتال , رفتار مشتري , توسعه محصول , هوش مصنوعي , ترجيح مشتري

innovation , genetic algorithm , digital innovation , customer behavior , Product development , artificial intelligence , customer preferences

Aigin Karimzade

Dr. Amir Zakeri