Loom output prediction method based on big data

Abstract

The present invention relates to the field of data processing, and in particular to a method for predicting loom production based on big data, the method comprising: obtaining observed values ​​and predicted values ​​of the loom's historical period production after preprocessing; calculating the loom's influencing factor, adjusting a preset smoothing coefficient based on the influencing factor to obtain an optimal smoothing coefficient, using the optimal smoothing coefficient to predict the loom's future period production to obtain a predicted value of the future period production; and adjusting the loom's production schedule based on the predicted production value. The present invention preliminarily evaluates the reliability of the historical period production prediction value, calculates an influencing factor based on the evaluation result, and introduces the influencing factor into exponential smoothing, so that the exponential smoothing model can adapt to the fluctuation of production, thereby improving the accuracy of the prediction and making more sensible decisions for the loom, thereby improving the production efficiency of the loom.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and more specifically, to a method for predicting loom output based on big data. Background Art

[0002] In the field of textile industry, due to the ever-changing market demand, enterprises are facing great production pressure and need more efficient production management to cope with it. Traditional management methods are inefficient and resource utilization is inadequate. Now, with the development of big data technology, textile enterprises can gain deeper insights by collecting and analyzing data from different sources (such as production line data, machine operation status, raw material information and market dynamics, etc.). This loom output forecasting method based on big data can help enterprises better predict market demand, optimize production plans, reduce waste, improve efficiency, and thus enhance the competitiveness of enterprises.

[0003] At present, the prior art, such as the patent application document CN115423223A, discloses an industrial Internet production prediction device and method for machinery manufacturing enterprises. The prediction method of the patent application document includes: the processor first determines the historical production volume of each statistical period based on the historical production volume data, and stores these data in a register. Then, the processor will call the historical production volume of each statistical period from the register and calculate the smoothing coefficient. Based on this smoothing coefficient, the processor will perform multiple exponential smoothing predictions on the historical production volume of each statistical period, obtain the exponential smoothing value of each statistical period, and store these values ​​in the second memory.

[0004] However, the historical production data collected has certain fluctuations, and the use of exponential smoothing prediction still leads to inaccurate prediction results, which reduces the production efficiency of the loom. Summary of the invention

[0005] In order to solve the technical problem that the historical production data collected above has certain fluctuations, and the use of exponential smoothing prediction still leads to inaccurate prediction results, thereby reducing the production efficiency of the loom, the present invention provides solutions in the following aspects.

[0006] The loom production prediction method based on big data includes:

[0007] Obtain the observed and predicted values ​​of the historical cycle output of the loom after preprocessing;

[0008] The preset smoothing coefficient is adjusted based on the influencing factor of the loom to obtain the optimal smoothing coefficient, and the future cycle output of the loom is predicted using the optimal smoothing coefficient to obtain the predicted value of the future cycle output;

[0009] Adjust the production schedule of the loom based on the output forecast value;

[0010] Among them, the impact factor is:

[0011] ; In the formula, is the impact factor, is an exponential function with the natural constant e as base, For the history The reference degree of the corresponding output forecast value of the sky loom, is the average reference value of the corresponding output forecast value of the loom in the historical period, The total number of historical days.

[0012] The present invention obtains the actual value and predicted value of the historical cycle output of the loom, preliminarily evaluates the reliability of the historical cycle output prediction value, calculates an impact factor based on the evaluation result, and introduces the impact factor into exponential smoothing to obtain the optimal smoothing coefficient, so that the exponential smoothing model can not only adapt to the fluctuation of output, but also achieve a balance between the responsiveness and stability of the prediction, thereby improving the accuracy of the prediction and making more informed decisions on the loom, thereby improving the production efficiency of the loom.

[0013] Preferably, the reference degree satisfies the relationship:

[0014] ; In the formula, For the history The reference degree of the corresponding output forecast value of the sky loom, For the history The predicted value of the corresponding output of the Sky Loom, For the history The observed value of the output of the Sky Loom, It is the maximum absolute value of all prediction errors corresponding to the loom in the historical period.

[0015] By calculating the reference degree of the predicted value of loom output in the historical period, it is possible to identify which days have more accurate predictions, thus providing a basis for subsequent improvements to the prediction model.

[0016] Preferably, the process of obtaining the optimal smoothing coefficient includes:

[0017] The product of the influencing factor and the preset smoothing coefficient is used as the optimal smoothing coefficient.

[0018] The optimal smoothing coefficient is determined by multiplying the influencing factor with the preset smoothing coefficient, so that the prediction model can dynamically adjust the predicted value according to the actual data, thereby more accurately reflecting future trends.

[0019] Preferably, the method of using the optimal smoothing coefficient to predict the future cycle output of the loom comprises:

[0020] In the prediction process, when the actual output of a new loom is generated, the smoothing coefficient corresponding to the actual output of the new loom is updated and iteratively calculated, and the updated smoothing coefficient is used to predict the output of the loom in future cycles.

[0021] During the forecasting process, as new data is continuously added, new smoothing coefficients are updated and iteratively calculated in real time. Using the updated smoothing coefficients for forecasting helps reduce forecast errors and improve forecast accuracy.

[0022] Preferably, the method of using the optimal smoothing coefficient to predict the future cycle output of the loom further includes:

[0023] When the number of iterations reaches the preset upper limit, the iteration process stops.

[0024] Preferably, the method of using the optimal smoothing coefficient to predict the future cycle output of the loom further includes:

[0025] During the iteration process, when the difference between the new smoothing coefficient and the last calculated smoothing coefficient is less than a preset threshold, the iteration is stopped.

[0026] Setting an upper limit on the number of iterations can prevent the model from overfitting on the training data and prevent the model from becoming too complex and losing its generalization ability. Similarly, when the new smoothing coefficient changes very little, it means that the model has stabilized, and further iterations may not bring significant performance improvements, but may lead to overfitting.

[0027] Preferably, the process of obtaining the reference degree includes:

[0028] Construct a box plot using the preprocessed predicted values ​​of the historical period output of the loom;

[0029] The distance between each predicted value and the corresponding quartile in the box plot is calculated, and the distance is used as the abnormality degree of the corresponding predicted value and normalized, and the normalized abnormality degree is used as the reference degree of the predicted value.

[0030] The box plot can be used to visually identify outliers in the data, thereby evaluating the accuracy of the predicted value. The quartiles in the box plot represent the central tendency of the data, and the distance between the predicted value and these quartiles reflects the degree of deviation of the predicted value. The smaller the deviation, the closer the predicted value is to the true value, and therefore has a higher reference value.

[0031] Preferably, the optimal smoothing coefficient satisfies the relationship:

[0032] ; In the formula, is the optimal coefficient, is the influencing factor of the loom, is the preset smoothing coefficient, are the preset parameters.

[0033] The beneficial effects of the present invention are:

[0034] The present invention analyzes the observed and predicted values ​​of the historical cycle output of the loom, takes into account the accuracy of the historical output prediction values, continuously optimizes the smoothing coefficient and performs iterative prediction, thereby achieving a more accurate prediction of the future output of the loom, and adjusts the production schedule accordingly, which helps to improve the production efficiency of the loom. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0036] Figure 1 It is a method flow chart of steps S1 to S3 in the loom output prediction method based on big data in an embodiment of the present invention. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0038] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0039] Reference Figure 1 The loom output prediction method based on big data includes steps S1 to S3, which are as follows:

[0040] S1: Obtain the observed and predicted values ​​of the historical cycle output of the loom after preprocessing.

[0041] In the embodiment of the present invention, it is set to collect the production data (ie, the observed value) of each loom after finishing work every day in a historical period, such as the past 100 days.

[0042] Next, taking a loom as an example, we use the exponential smoothing method to predict the production data of the loom in the historical period. The specific steps are as follows:

[0043] The first step is to set a smoothing coefficient. In the embodiment of the present invention, the smoothing coefficient is set to 0.7. In other embodiments, the implementer can adjust the value of the smoothing coefficient according to the actual situation.

[0044] The second step is to start from the first observation in the historical period and use a fixed smoothing coefficient to predict the observed values ​​of the loom.

[0045] Step 3: Calculate the predicted value; the predicted value satisfies the following relationship:

[0046]

[0047] In the formula, For the history The predicted value of the corresponding output of the Sky Loom, For the history The observed value of the output of the Sky Loom, For the history The predicted value of the corresponding output of the Sky Loom, is the smoothing coefficient set above.

[0048] Repeat the first to third steps above to obtain the predicted value of the loom's daily output data within the historical period.

[0049] After obtaining the observed and predicted values ​​of the historical cycle output of the loom, the collected observed and predicted values ​​are cleaned and preprocessed, including removing outliers, filling missing values, and data standardization to ensure the quality and consistency of the data.

[0050] S2: Based on the influencing factors of the loom, the preset smoothing coefficient is adjusted to obtain the optimal smoothing coefficient, and the future cycle output of the loom is predicted using the optimal smoothing coefficient to obtain the predicted value of the future cycle output.

[0051] In traditional exponential smoothing methods, a fixed smoothing coefficient is usually used to weight all historical data. This means that no matter how far the data points are from the forecast point, their impact on the forecast result is consistent. However, in actual situations, data sequences are constantly changing, and some data points may be more important or more reflective of future trends than other data points. Since a single smoothing coefficient cannot adapt to this changing reference degree, it may not accurately reflect the actual importance of the data points, resulting in inaccurate forecast results. In order to solve this problem, the following analysis is conducted.

[0052] The accuracy of the prediction (or the difference between the predicted value and the actual value) directly affects the reference value of the predicted value for subsequent predictions. Simply put, if a prediction is inaccurate, then its reference value for similar predictions in the future is low. The reference value of each data point for subsequent predictions can be calculated based on the difference between the predicted value and the actual value.

[0053] Exemplarily, the reference degree satisfies the relationship:

[0054]

[0055] In the formula, For the history The reference degree of the corresponding output forecast value of the sky loom, For the history The predicted value of the corresponding output of the Sky Loom, For the history The observed value of the output of the Sky Loom, It is the maximum absolute value of all prediction errors corresponding to the loom in the historical period.

[0056] According to the history The reference degree of the output forecast value corresponding to the Tianzhi Machine can be obtained in the same way as the reference degree of all forecast values ​​in the historical period.

[0057] Among them, the history The reference degree of the corresponding output forecast value of the sky loom As an indicator reflecting the accuracy of the forecast, its value is between 0 and 1. The closer it is to 1, the more historical The more accurate the forecast is.

[0058] In another embodiment, another method is provided to obtain the above-mentioned reference degree. The general steps are as follows:

[0059] Construct a box plot using the preprocessed predicted values ​​of the historical period output of the loom;

[0060] The distance between each predicted value and the corresponding quartile in the box plot is calculated, and the distance is used as the abnormality degree of the corresponding predicted value and normalized so that the abnormality degree value is between 0 and 1. The normalized abnormality degree is used as the reference degree of the predicted value.

[0061] Among them, if a predicted value is very close to the quartile (i.e., the degree of abnormality is low), its referenceability is high; if a predicted value is far away from the quartile (i.e., the degree of abnormality is high), its referenceability is low.

[0062] For any loom, its referenceability is an important indicator to measure the reliability of its production forecast value. The lower the referenceability, the less reliable the production forecast value of the loom is, so the data should be considered less in the subsequent loom production forecast process.

[0063] There may be differences in reliability between the output forecasts of different looms or the same loom at different points in time. This difference may be due to a variety of factors, such as equipment status, raw material quality, operator proficiency, etc. Therefore, a method is needed to quantify this reliability difference so that more reliable data can be given greater weight in the forecast process. In addition, in loom output forecasting, recent data is generally considered to be more valuable than historical data. This is because recent data can better reflect the current status and trends of the loom. By calculating the impact factor, more weight can be given to recent data, thereby more accurately reflecting the current situation of the loom. The smoothing coefficient is an important parameter in the forecasting model, which determines the proportion of historical data and current data in the forecast result. By calculating the impact factor, the smoothing coefficient can be adaptively adjusted according to the reliability difference and time sensitivity of the data.

[0064] Exemplarily, the influence factor of the loom satisfies the relationship:

[0065]

[0066] In the formula, is the influencing factor of the loom, is an exponential function with the natural constant e as base, For the history The reference degree of the corresponding output forecast value of the sky loom, is the average reference value of the corresponding output forecast value of the loom in the historical period, The total number of historical days.

[0067] The above process of calculating the impact factor is applied to all looms to obtain the impact factor corresponding to each loom.

[0068] After calculating the influence factors of each loom, the smoothing coefficient can be adaptively adjusted according to these factors. For example, for looms with higher influence factors, their predicted values ​​can be given more weight; for looms with lower influence factors, their predicted values ​​can be given less weight.

[0069] The size of the smoothing coefficient directly affects the sensitivity and stability of the prediction results. When there are outliers or noise in the data, the fixed smoothing coefficient may be greatly affected, resulting in unstable prediction results. Therefore, by calculating the influencing factor of each loom and adaptively obtaining its smoothing coefficient based on the influencing factor, the prediction of loom production data is further made more accurate and reliable.

[0070] Exemplarily, the product of the influence factor calculated above and the smoothing coefficient preset in step S1 is used as the optimal smoothing coefficient. The optimal smoothing coefficient satisfies the relationship:

[0071]

[0072] In the formula, is the optimal coefficient, is the influencing factor of the loom, is the preset smoothing coefficient.

[0073] In other embodiments, the optimal smoothing coefficient also satisfies the relationship:

[0074]

[0075] In the formula, is the optimal coefficient, is the influencing factor of the loom, is the preset smoothing coefficient, are the preset parameters.

[0076] Then the optimal smoothing coefficients corresponding to all looms are calculated.

[0077] Furthermore, the optimal smoothing coefficient and exponential smoothing formula are used to predict the future cycle output based on the historical cycle loom output data. The specific cycle can be set according to the forecast demand or data availability.

[0078] It should be noted that in the prediction of loom production, data is constantly changing over time. New data may contain factors or trend changes that were not previously considered. Therefore, with the addition of new data, iterative updates can gradually optimize the smoothing coefficient to make it more consistent with the actual situation of the data.

[0079] For example, after obtaining new time-series loom production data (i.e., new actual values), it is necessary to iteratively calculate a new smoothing coefficient. The new smoothing coefficient will be calculated based on the new actual value and the previous predicted value to reflect the latest trend of the data. The updated smoothing coefficient will be used to continue predicting the loom production data to obtain a more accurate predicted value.

[0080] When the number of iterations reaches the preset upper limit (the specific setting can be determined through experiments and verification to determine the optimal upper limit of the number of iterations), the iteration process is stopped. This preset upper limit is to prevent the iteration process from going on indefinitely and wasting computing resources, and also to avoid problems such as overfitting or oscillation. After the iteration is stopped, the smoothing coefficient obtained from the last iteration will be used to predict future production.

[0081] In another embodiment, when the difference between the new smoothing coefficient and the last calculated smoothing coefficient is smaller than a preset threshold, the iteration is stopped.

[0082] Then the output forecast values ​​of all looms in the future cycle are predicted.

[0083] S3: Adjust the production schedule of the loom based on the output forecast value.

[0084] According to the output forecast value of the loom in the future cycle obtained in step S2, the production process is optimized to ensure that each loom can operate in an efficient and stable state. For example, the operation time, maintenance plan and staffing of the loom can be adjusted.

[0085] Alternatively, if the forecast output is higher than expected, the production task of the loom can be increased; if the forecast output is lower than expected, the production task can be reduced or reallocated.

[0086] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0087] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A loom production prediction method based on big data, characterized in that: include: Obtain the observed and predicted values ​​of the historical cycle output of the loom after preprocessing; The preset smoothing coefficient is adjusted based on the influencing factor of the loom to obtain the optimal smoothing coefficient, and the future cycle output of the loom is predicted using the optimal smoothing coefficient to obtain the predicted value of the future cycle output; Adjust the production schedule of the loom based on the output forecast value; Among them, the impact factor is: ; In the formula, is the impact factor, is an exponential function with the natural constant e as base, For the history The reference degree of the corresponding output forecast value of the sky loom, is the average reference value of the corresponding output forecast value of the loom in the historical period, is the total number of historical days; The reference degree satisfies the relationship: ; In the formula, For the history The predicted value of the corresponding output of the Sky Loom, For the history The observed value of the output of the Sky Loom, It is the maximum absolute value of all prediction errors corresponding to the loom in the historical period.

2. The method for predicting loom output based on big data according to claim 1, characterized in that: The process of obtaining the optimal smoothing coefficient includes: The product of the influencing factor and the preset smoothing coefficient is used as the optimal smoothing coefficient.

3. The method for predicting loom output based on big data according to claim 2 is characterized in that: The method of using the optimal smoothing coefficient to predict the future cycle output of the loom includes: In the prediction process, when the actual output of a new loom is generated, the smoothing coefficient corresponding to the actual output of the new loom is updated and iteratively calculated, and the updated smoothing coefficient is used to predict the output of the loom in future cycles.

4. The method for predicting loom output based on big data according to claim 3 is characterized in that: The use of the optimal smoothing coefficient to predict the future cycle output of the loom also includes: When the number of iterations reaches the preset upper limit, the iteration process stops.

5. The method for predicting loom output based on big data according to claim 3 is characterized in that: The use of the optimal smoothing coefficient to predict the future cycle output of the loom also includes: During the iteration process, when the difference between the new smoothing coefficient and the last calculated smoothing coefficient is less than a preset threshold, the iteration is stopped.

6. The method for predicting loom output based on big data according to claim 1, characterized in that: Or the process of obtaining the reference degree includes: Construct a box plot using the preprocessed predicted values ​​of the historical period output of the loom; The distance between each predicted value and the corresponding quartile in the box plot is calculated, and the distance is used as the abnormality degree of the corresponding predicted value and normalized, and the normalized abnormality degree is used as the reference degree of the predicted value.

7. The method for predicting loom output based on big data according to claim 1, characterized in that: The optimal smoothing coefficient satisfies the relationship: ; In the formula, is the optimal coefficient, is the influencing factor of the loom, is the preset smoothing coefficient, are the preset parameters.

Images