In addition to Weibo, there is also WeChat
Please pay attention
WeChat public account
Shulou
2025-01-19 Update From: SLTechnology News&Howtos shulou NAV: SLTechnology News&Howtos > Internet Technology >
Share
Shulou(Shulou.com)06/03 Report--
For things that are uncertain in the future, everyone is used to using past experience to predict. In this era of rapid development, is it accurate to rely on experience to predict?
Prediction is what is often called pre-speculation and determination. its purpose is not to predict for prediction, but to guide human behavior decision-making, so that people will not worry about the uncertainty of the future when making decisions. Many people often speculate on the basis of experience. Although this method is simple and convenient, it lacks theoretical basis and is subjective. The end result is that the forecast reflects personal will, not reality.
Avenash Kosik (Avinash Kaushik), an expert in digital marketing and analysis, once wrote in his blog Occam Razor: "you and I are wrong 80 per cent of the time in predicting consumer demand."
Unfortunately, those who manipulate statistical benchmark forecasts feel that they can improve the accuracy of predictions by using their own judgment. Through the research, it is found that when forecasters improve their predictions, they are almost always wrong, because they are too optimistic, which leads to lower prediction accuracy. On the contrary, when forecasters reduce the forecast value, because it is more conservative, it can often improve the prediction accuracy. In general, the subtle changes that affect the rise or fall of the forecast have little effect on the prediction accuracy, which is a waste of time. So how to use scientific methods to predict it?
We can use business knowledge to predict the future based on data and analysis. With the development of computer technology and network technology, big data's technology has penetrated into various industries. It is an inevitable trend to mine potentially valuable relationships, trends and patterns from massive data, to build forecasting models and to make predictive analysis. Forecasting through data can help enterprises to find market opportunities and make scientific business decisions.
Scientific prediction is inseparable from data, data can not be separated from forecasting methods, the current forecasting methods are roughly divided into the following categories:
Qualitative prediction method
It mainly depends on people's subjective judgment. When there is little historical data available for reference, the qualitative prediction method is the most appropriate.
Time series prediction method
Using historical data to predict the future, it is especially suitable for scenarios where the basic pattern does not change much from year to year.
Causality prediction method
Assuming that demand forecasting is related to some factors, causality forecasting method can find the relationship between these factors and demand, and predict the future by predicting the changes of these external factors.
Simulation method
The simulation model allows forecasters to make certain assumptions about the predicted conditions.
Scientific forecasting needs to go through determining demand-obtaining data-analyzing data-building models-predicting the future-supporting decision-making.
The first step is to determine the object, goal and scope of the prediction, which includes geographical scope and time range. Collect the required data, preprocess the data, and analyze the periodicity, seasonality, trend and randomness of the data. Select the prediction method to establish the model, and at the same time confirm whether the model is effective for prediction. Make a reasonable prediction to the forecast object according to the previous data information and prediction model. Through the prediction results, decisions can be made for the things to come in order to achieve the predicted goals.
Next, let's take a simple example. If you are asked to predict the sales volume of a company's product in the next few months, and these predicted variables have a growth trend, the company may make strategic adjustments and layouts based on your forecast results. so how to predict? Time series prediction can be used.
The sample data are as follows:
Explore and analyze the sales data, the distribution map of monthly sales volume of goods:
Distribution map of sales volume of goods in each province:
Distribution map of sales volume of commodity models in each province:
The basic information of data obtained by using Python is as follows:
It can be seen from the above that the amount of data is 218618, and there is no missing value, so there is no need to deal with the missing value. Next, we need to count the monthly product sales, and then select the time series method to predict. Before the time series analysis, we need to analyze the stability of the data, and use the difference method to stabilize the unstable data. At the same time, the data is decomposed to analyze the periodicity, seasonality, trend and randomness of the data. I will not repeat the details here. This paper chooses the method of SARIMAX based on python. SARIMAX is based on the differential integrated moving average autoregressive model (ARIMA) by adding seasonal, periodic and eXogenous external factors of Seasonal. The generated model is summarized as follows:
Using the model to predict the results in the next few months is as follows:
Concluding remarks
In the current era of digital economy, self-righteousness is bound to fail. Use data with business knowledge, using proven analysis, rather than relying on pure intuition. Use scientific methods to analyze and foresee its development trend, grasp the law of market changes, improve the scientific level of management, reduce the blindness of decision-making, reduce the uncertainty in the future, and reduce the risks that may be encountered in decision-making. only in this way can the decision-making goal be realized smoothly.
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: 0
*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.