Werewolf game is one of the popular party games. People enjoy not only playing but also watching games. Besides, Werewolf game is studied as an incomplete information game. In this paper, the authors focused on the watching support for Werewolf game. A predictor built with machine learning (logistic regression) calculates the situation of the game by inputting events that occur in the game and provides each player’s execution/attack probability to the audience. Though the predictor was originally built to analyze the games, it is applied to watching support in this paper. In advance, the predictor had to be constructed by analyzing game logs played in the target regulation. In this paper, the authors focused on five-player games as the fundamental setting. The sets of game logs in different situations were inputted in parts by parts. The predictor appropriately estimated each player’s execution/attack probability in most cases. The estimation will help the audience watch Werewolf games. Applying this method to different regulations is one of the future works.

This paper aims to sustain the liveliness of karaoke parties. To do this, the authors focused on programs for pop music concerts. Since they are carefully organized to make the concerts lively, they should also be useful to make karaoke parties lively. The authors’ principal idea is to reorder the requested songs at the karaoke party to follow the inclination of the order of songs in the concerts. As a preparation, the authors collected the programs of pop music concerts and added feature vectors to each song. In the parties, the score of each requested song is calculated with the feature vector. This score is higher when the song is expected to be performed at that time in the concerts. Based on the score and consideration of the participants’ waiting time, the songs in the request queue are reordered. The song orders like the music concert are expected to be reproduced at the karaoke parties. In the test drive, the prototype system worked as expected and at least didn’t disturb the karaoke experience. Development of the demonstrative system and tuning the algorithm are the major future works.