Modern technology innovations feature a successive and even recurrent procedure. Intervals between old and new generations of technology are shrinking, and the Internet and Web services have facilitated the fast adoption of an innovation even before the convergence of its predecessor. While the adoption and diffusion of innovations have been studied for decades, most theories and analyses focus on single and one-time innovations. Meanwhile, limited work has investigated successive innovations while lacking user-level analysis, possibly due to the unavailability of fine-grained adoption behavior data. In this study, we present the first large-scale analysis of the adoption of recurrent innovations in the context of mobile app updates, investigating how millions of users consume various versions of thousands of apps on their mobile devices. Our analysis reveals novel patterns of crowd and individual adoption behaviors, which suggest the need for new categories of adopters to be added on top of the Rogers model of innovation diffusion. We show that standard machine learning models are able to pick up various sources of signals to predict whether users in these different categories will adopt a new version of an app and how soon they will adopt it.

Depression is a serious psychiatric disorder that is prevalent worldwide and is usually characterized by persistently depressed mood, impaired mobility, and delayed thinking and cognitive functions. This experiment uses the Kinect V2 device to record simple kinematic skeletal data of body joints of depressed patients and non-depressed patients, directly extracting the presented spatial features and low-level features from the recorded raw Kinect-3D coordinates. Aiming at the symptoms of delayed thinking and cognitive function and impaired mobility in patients with depression, the features of reaction time are obtained from preprocessing, and this prior knowledge is added to the deep learning model to assist the recognition and classification of the model, thereby improving the classification accuracy. The objective of this project is to develop a deep learning model for detecting depression using preprocessed data.

Depression is a common mental illness. Unlike normal mood fluctuations which affect individuals only temporarily, depressed episodes can profoundly disrupt a person’s daily life and even lead to suicide. Eye movement data are commonly employed in depression identification because they are simple to collect and can show psychological processes. Given the above, we proposed EnSA, a novel model based on eye movement data. We established forward and reverse target stimuli to identify the subject’s saccade reaction and capture the subject’s eye movement data. The gathered eye movement data were entered into EnSA first, and the self-attention weights of each characteristic were calculated. To produce more expressive features, the self-attention features were convolved and then summed feature outputs. According to the results, our model performed well, with an accuracy of 93.5% and 95.5% in the prosaccade and antisaccade experiments, respectively.

Eye movement trajectories of depressed patients and normal persons are different. The eye-tracking data obtained by the eye tracker can adequately summarize the characteristics of the eye movement trajectory. Based on the characteristics of eye movement trajectory, this paper proposes a new depression detection model by using an artificial neural network, which can better assist doctors in the diagnosis of depression. First, we extract the feature of eye movement trajectory, which obtains from time-series data recording the trajectory of the eye. Then, we convert the data from three-dimensional to two-dimensional, and perform feature extraction and transformation. Finally, we propose a new depression detection model by using artificial neural networks. The experimental results show that the best result of the model evaluation is 83.17%, which can effectively assist doctors in the diagnosis of depression.