DS Note to Self - 032819 - The 5 Basic steps in forecasting tasks 預測事情的五個基本步驟

photo from from pixabay

Forecasting sounds like a magic word. However, it is required in many situations such as stocking an inventory, planning for retirement, or even budgeting for a trip. It is necessary for our daily life. Forecasting is an important aid to effective and efficient planning.

Below, I will list out 5 general steps to take for a forecasting job.

Step 1. Problem Definition

This is often the most difficult step of forecasting. Simply because you don't know what you don't know if you are not good at observing problems. Defining problems requires an understanding of the problem itself. And what forecast style fit the organization to use based on different skill sets of the members have or required.

Step 2. Gathering Information

One needs to know what information needs to be collected, says, how the way and what kind of the data set is collected. And how the data set going to be stored as for future usage. There are always at least two kinds of information you can encounter with when dealing with forecasting setup:

1. Statistical data
2. The accumulated experience of the people who carrying the forecasting.

Sometime, it would be difficult to obtain enough historical data to fit a statistical model. You can try judgmental forecasting method if so. Not all historical data are useful, it depends the requirement and scope of the forecasting project.

Step 3. Exploratory Analysis

If you have collect some amount of data and wanted to start get to know more about the data. You can start by graphing the data. You can check if there are any consistent patterns such as if any significant trend, seasonality and cycles. Are there any outliers that can be explained. There are various ways to help with the analysis.

Step 4. Selecting and Fitting Models

Once you see any special patterns, you can try to find different tools that have been developed by people and fit the data in and then adjusting the variable parameters of it to make it more usable for your data set.

Step 5. Applying and Evaluating the Forecasting Models

Once you found a suitable statistical forecasting model. You can evaluate the performance of the model by running with different variable parameters. Many tools online have been developed to help improving the accuracy of the forecasting models.

So, there you have it, these are the five basic steps in forecasting tasks.

---

預測聽起來像一個神奇的詞。但是，在許多情況下都需要它，例如存貨，計劃退休，甚至預算旅行。這對我們的日常生活是必要的。預測是有效和高效規劃的重要輔助手段。

下面，我將列出預測工作的5個一般步驟。

第1步。問題定義

這通常是預測中最困難的一步。只是因為你不知道你不知道你不擅長觀察問題。定義問題需要了解問題本身。什麼預測風格適合組織根據成員的不同技能組合使用或要求。

第2步。收集信息

人們需要知道需要收集哪些信息，如何收集數據集的方式和類型。以及如何存儲數據集以供將來使用。在處理預測設置時，您始終可以遇到至少兩種信息：

1）統計數據
2）進行預測的人的累積經驗。

有時，很難獲得足夠的歷史數據來擬合統計模型。如果是這樣，您可以嘗試判斷預測方法。並非所有歷史數據都有用，它取決於預測項目的要求和範圍。

第3步。探索性分析

如果您已收集了一些數據並希望開始了解有關數據的更多信息。您可以從繪製數據圖表開始。您可以檢查是否存在任何一致的模式，例如是否有任何重要趨勢，季節性和周期。有沒有可以解釋的異常值。有多種方法可以幫助進行分析。

步驟4.選擇和擬合模型

一旦看到任何特殊模式，您就可以嘗試找到人們開發的不同工具並將數據擬合，然後調整它的變量參數，使其更適用於您的數據集。

步驟5.應用和評估預測模型

一旦找到合適的統計預測模型。您可以通過運行不同的變量參數來評估模型的性能。已經開發了許多在線工具來幫助提高預測模型的準確性。

這是預測事情的五個基本步驟。

---

_{Sponsored ( Powered by dclick )}

WANT TO EARN CRYPTOCURRENCY FROM HOME WITHOUT INVESTMENT?

^{CHECK OUT MY REVIEW OF BOUNTY0X.}