In addition to Weibo, there is also WeChat
Please pay attention
WeChat public account
Shulou
2025-01-23 Update From: SLTechnology News&Howtos shulou NAV: SLTechnology News&Howtos > Internet Technology >
Share
Shulou(Shulou.com)06/03 Report--
Number is a product of Getui that provides statistical analysis of data to APP developers. Through visual embedding technology and big data's analysis ability, it makes a comprehensive statistical analysis of APP from the dimensions of user attributes, channel quality and industry comparison.
"number" can not only count the active and new users in time, but also analyze the composition and flow direction of unloaded users, and predict the key behaviors of users, such as loss and payment, so as to help APP developers achieve fine operation and lifecycle management of users. It is worth mentioning that the innovation of "number" in "visual embedding" and "behavior prediction" has brought great convenience for APP developers in actual operation, so, in the following, we will also do a detailed analysis around these two points.
Visual burial point
Burying point refers to the process of embedding relevant statistical codes in key parts of the product process to track user behavior, count the usage of key processes, and report the data to the server in the form of logs.
At present, the data burying point acquisition mode mainly includes code embedding point, no burying point, visual burying point and so on.
"Code burying point" refers to adding basic js to the monitoring page and adding monitoring code according to the requirements. Its advantage is flexibility, you can customize settings, you can choose the data you need to analyze, but for complex websites, every time you modify a page, you have to make a new buried point plan, which is costly. At present, Baidu Statistics, Friends, Tencent Cloud Analytics, Google Analytics and other representative products have been adopted.
"Visualization embedding" usually means that developers connect user behavior analysis tools through devices, directly operate interactive and effective page elements (such as pictures, buttons, links, etc.) on the data access management interface to achieve data burial, and send the collection code effective return number. At present, the representative products of visual embedding sites are the number of products, Mixpanel, Shenze data and so on.
"No burying point" is similar to "full buried point", its principle is "collect all, select on demand", that is to say, it can collect the user behavior of all interactive elements in the page, it first collects and detects the content of the page as much as possible, and then determines which data to analyze through the interface configuration, but it is a standardized collection, if you need to set a custom collection method, you still need code embedding to help. The representative products of this scheme are GrowingIO, several geeks, Baidu Statistics and so on.
Why did the "number" choose the visual burying point?
At present, the mobile Internet is in a stage of rapid development and rapidly changing development situation, developers need to make timely adjustments to business functions according to big data's analysis and feedback, in the traditional mode of operation, if you want to understand the data of different nodes, you have to modify the buried points in the corresponding code, then test and release, and then review and launch in the app store, the whole cycle may be as long as several weeks. This obviously does not meet the needs of the business. Therefore, the "visual embedding" technology adopted by "number" is to help developers solve this problem.
The visual embedding point of "number" is flexible and convenient, and there is no need to add any code to the data tracking point. Users only need to connect to the management desk through the device, circle the elements that can be buried on the page, and then add interface tracking points that take effect at any time. At the same time, in terms of data collection mode and data analysis ability, "number" can provide developers with accurate and effective data.
Visual burial points mainly have the following characteristics:
1. Zero code, no code, cost saving
2. Update-free, convenient to add and no need to upgrade
3. Easy to test, circle test, real-time presentation
In other words, visual burial sites can not only save enterprise costs, but also improve the productivity of developers and operators.
Behavior prediction
The behavior prediction of "number" mainly includes loss prediction, uninstall prediction, payment prediction and so on. Its principle is to build an algorithm model based on App historical behavior data to predict users' key behavior, so as to help developers achieve the purpose of refined operation and life cycle management.
It should be noted here that the behavior prediction of "number" is different from the personalized recommendation commonly used in e-commerce platform, the latter is mainly based on users' recent behavior, such as browsing records and purchase records, and analyzes what users may need, while "number" is based on the comprehensive analysis of the number and trend of unloading in each channel of App, and is more based on the cluster analysis of the crowd, rather than just based on individual behavior.
Steps to predict behavior
According to Zhu Jinxing, a scientist with "Getui" big data, the behavior prediction of "number" is mainly divided into the following steps:
1. Find samples, mainly from the historical database
2. feature extraction to connect the user with the database and make a match.
3. Feature screening to retain highly relevant or valuable features
4. In the model training, the retained features are put into the model for training. In the selection of the model, the "number" mainly uses logical regression, logical regression model and other models, which are relatively simple and relatively easy to deal with in feature screening, and the results are easy to explain and relatively stable.
5. Optimize the parameters and adjust them according to the effect. If the result is not satisfactory, you can return to adjust the parameters and go back to the process more than once.
Case analysis
Let's take the payment forecast as an example to sort out the specific implementation process.
The process of forecasting individual payment mainly includes the following points:
1. Target problem decomposition
Identify the issues that need to be predicted, that is, paid forecasts, and the span of time ahead.
2. Analyze the sample data
(1) extract the historical payment records of all users
(2) analyze payment records to understand the composition of paying users, such as age, sex, purchasing power and product categories consumed.
(3) extract the historical data of non-paying users. Here, according to the needs of the product, you can add conditions or extract unconditionally, such as extracting active and non-paying users, or extracting directly without conditions.
(4) analyze the composition of non-paying users.
3. The characteristics of building the model.
(1) the original data may be used directly as features.
(2) some data can be used better after transformation, such as age, which can be transformed into juvenile, middle-aged, old-aged and other characteristics.
(3) the generation of cross features, such as "middle-aged" and "female", can be combined into one feature.
4. Calculate the correlation of features.
(1) calculate the feature saturation and filter the saturation
(2) calculate the feature IV, Chi-square and other indexes to filter the feature correlation.
5. Logical regression is selected for modeling.
(1) choose the appropriate parameters for modeling.
(2) after the model is trained, the accuracy, recall rate and AUC of the model are calculated to evaluate the model.
(3) if you think the performance of the model is acceptable, you can do verification on the verification set, and after the verification is passed, the model is saved and predicted.
6. Forecast
Load the above saved model, and load the prediction data for prediction.
7. Monitoring
Finally, operators also need to monitor the key indicators of each forecast, identify and solve problems in a timely manner, and prevent unexpected situations, resulting in invalid forecasts or deviations in prediction results.
Other scenarios, such as loss prediction, uninstall prediction, etc., are similar to payment prediction in process, so I will not cover them one by one here.
With accurate behavior prediction, operators can split and refine the operational objectives, specific to each scenario, each process, and adopt different promotion channels and operational strategies for different users. For example, based on churn prediction, operators can gain insight into user churn behavior in advance, intervene in advance, and retain the soon-to-be-lost users through personalized content recommendation, message push and other operational means, so as to reduce the turnover rate. Generally speaking, with the help of big data's behavior prediction, operators can understand users more timely and comprehensively, so as to achieve the goal of fine operation.
About the future.
Next, "number" will also do more exploration in commodity recommendation and other areas, such as the development of accurate recommendation technology, and will also continue to tap big data's potential and make further optimization combined with feedback data. do more in-depth training and study around the sample data provided by customers, to provide developers with more comprehensive big data services, please look forward to.
Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.
Views: 207
*The comments in the above article only represent the author's personal views and do not represent the views and positions of this website. If you have more insights, please feel free to contribute and share.
Continue with the installation of the previous hadoop.First, install zookooper1. Decompress zookoope
"Every 5-10 years, there's a rare product, a really special, very unusual product that's the most un
© 2024 shulou.com SLNews company. All rights reserved.